All notable changes to this project will be documented in this file. This project adheres to Semantic Versioning.
This release includes a number of documentation fixes. It also includes the experimental binary for darwin/arm64.
There are no code changes.
- OpenTelemetry: fix configuration example, authored by @rvalkenaers
- Configuration: fix typo for
tls-cert-refresh-period
, authored by @mattmahn - SSH and Sudo authorization: Add missing filename
- Integration: fix example policy
- Build darwin/arm64 in post tag workflow
This release contains a number of fixes and enhancements.
This release adds OpenTelemetry support to OPA. This makes it possible to emit spans to an OpenTelemetry collector via gRPC on both incoming and outgoing (i.e. http.send) calls in the server. See the updated docs on monitoring for more information and configuration options (#1469) authored by @rvalkenaers
This release also adds a new opa exec
command for doing one-off evaluations of policy against input similar to
opa eval
, but using the full capabilities of the server (config file, plugins, etc). This is particularly useful in
contexts such as CI/CD or when enforcing policy for infrastructure as code, where one might want to run OPA with remote
bundles and decision logs but without having a running server. See the updated docs on
Terraform for an example use case.
(#3525)
- Four new functions for working with HMAC (
crypto.hmac.md5
,crypto.hmac.sha1
,crypto.hmac.sha256
, andcrypto.hmac.sha512
) was added (#1740) reported by @jshaw86 array.reverse(array)
andstrings.reverse(string)
was added for reversing arrays and strings (#3736) authored by @kristiansvalland and @olamiko- The
http.send
built-in function now uses a metric for counting inter-query cache hits (#4023) authored by @mirayadav - An overflow issue with dates very far in the future has been fixed in the
time.*
built-in functions (#4098) reported by @morgante
- A problem with future keyword import of
in
was fixed foropa fmt
(#4111) reported by @keshavprasadms - An issue with
opa fmt
when refs contained operators was fixed (authored by @jaspervdj-luminal) - Fix file renaming check in optimization using
opa build
(authored by @davidmarne-wf) - The
allow_net
capability was added, allowing setting limits on what hosts can be reached in built-ins likehttp.send
andnet.lookup_ip_addr
(#3665)
- A new credential provider for AWS credential files was added (#2786) reported by @rgueldem
- The new
--tls-cert-refresh-period
flag can now be provided toopa run
. If used with a positive duration, such as "5m" (5 minutes), "24h", etc, the server will track the certificate and key files' contents. When their content changes, the certificates will be reloaded (#2500) reported by @patoarvizu - A new
v1/status
endpoint was added, providing the same data as the status plugin would send to a remote endpoint (#4089) - The HTTP router of OPA is now exposed to the plugin manager (#2777) authored by @bhoriuchi reported by @mneil
- Calling
print
now works in decision masking policies - An unintended switch between long/regular polling on 304 HTTP status was fixed (#3923) authored by @floriangasc
- The error message about prohibited config in the discovery plugin has been improved
- The discovery plugin no longer panics in Trigger() if downloader is nil
- The bundle plugin now ignores service errors for file:// resources
- The bundle plugin file loader was updated to support directories
- A timer to HTTP request was added to the downloader
- The requested_by field in the logging plugin is now optional
-
The error message raised when using
-
with a number and a set is now more specific (as opposed to the correct usage with two sets, or two numbers) (#1643) -
Fixed an edge case when using print and arrays in unification (#4078)
-
Improved performance of some array operations by caching an array's groundness bit (#3679)
-
⚠️ Stricter check of arity in undefined function stage (#4054). This change will fail evaluation in some unusual cases where it previously would succeed, but these policies should be very uncommon.An example policy that previously would succeed but no longer will (wrong arity):
package policy
default p = false
p {
x := is_blue()
input.bar[x]
}
is_blue(fruit) = y { # doesn't use fruit
y := input.foo
}
- The
opa.runtime()
built-in is now made available to the SDK (#4050 authored by @oren-zohar and @cmschuetz - Plugins are now exposed on the SDK object
- The SDK now supports graceful shutdown (#3980) reported by @brianchhun-chime
print
output is now sent to the configured logger
- All pages in the docs now have a feedback button (#3664) authored by @alan-ma
- The Kafka docs have been updated to use the new Kafka plugin, and to use the OPA management APIs
- The Terraform tutorial was updated to use
opa exec
(#3965) - The docs on Contributing as well as the Vendor Guidelines have been updated
- The term "whitelist" has been replaced by "allowlist" across the docs
- A simple destructuring assignment example was added to the docs
- The docs have been reviewed on the use of assignment, equality and comparison operators, to make sure they follow best practice
- SHA256 checksums of CI builds now published to release directory (#3448) authored by @johanneslarsson reported by @raesene
- golangci-lint upgraded to v1.43.0 (authored by @shuheiktgw)
- The build now creates an executable for darwin/arm64. This should work as expected, but is currently tested in the CI pipeline like the other binaries
- PRs targeting the ecosystem page are now checked for mistakes using Rego policies
This release contains a number of fixes and enhancements.
This release adds an early exit optimization to the evaluator. With this optimization, the evaluator stops evaluating rules when an answer has been found and subsequent evaluation would not yield any new answers. The optimization is automatically applied to complete rules and functions that meet specific requirements. For more information see the Early Exit in Rule Evaluation section in the docs. #2092
- The
net.lookup_ip_addr
function was added to allow policies to resolve hostnames to IPv4/IPv6 addresses (#3993) - The
http.send
function has been improved to close TCP connections quickly after receiving the HTTP response and avoid creating HTTP clients unnecessarily when a cached response exists (#4015). This change reduces the number of open file descriptors required in high-throughput environments and prevents OPA from encountering ulimit errors.
print()
calls in the head of rules no longer cause runtime errors (#3967)- Type errors for calls to undefined functions no longer contain rewritten variable names (#4031)
- The
rego.SkipPartialNamespace
option now correctly sets the flag on the partial evaluation queries (previously it would always set the value totrue
) (#3996) authored by @thomascoquet - The internal set implementation has been updated to insert elements in sorted order rather than lazily sorting during comparisons.
- Fixed
import
alias parsing bug identified by fuzzer (#3988)
- The Golang SDK will now issue a
grow()
call if theinput
document exceeds the available memory space. - The
malloc()
implementation will now callopa_abort
if thegrow()
call fails.
- The decision logger adapts upload chunk sizes based on previous outputs. This allows the decision loggger to encode significantly more decisions into each upload chunk, thereby reducing heap usage for buffered decisions. For more information on the adapative chunking behaviour, see the Decision Logs page in the docs.
- The decision logger can be configured to send records to a custom plugin as well as an HTTP endpoint at the same time (#4013)
print()
calls from thesystem.authz
policy are now included in the logs (#4048)- OPA can use an Azure Managed Identities Token to authenticate with control plane services (#3916) authored by @Scowluga.
- The logging configuration will be correctly applied to service clients so that DEBUG logs are surfaced (#4071)
- The
opa fmt
command will not generate a line-break when there are generated variables in a function call (#4018) reported by @torsrex - The
opa inspect
command no longer prints a blank namespace when a data.json file is included at the root (#4022) - The
opa build
command will output debug messages if an optimized entrypoint is discarded.
- The website has been updated to build with Hugo 0.88.1 (#3787)
- The version picker in the documentation is now scrollable (#3955) authored by @orweis
- The description of the
urlquery
built-in functions have been clarified (#1592) reported by @klarose - The decision logger documentation has been improved to cover controls for large-scale environments (#3976)
- The "strict built-in errors" mode is now covered in the docs along with built-in function error behaviour (#3686)
- The OAuth2 and OIDC examples around key rotation and caching have been improved
- Issues and PRs that have not seen activity in 30 days will be automatically marked as "inactive"
- The
Makefile
can now produce Docker images for other architectures. We do not yet publish binaries or images for non-amd64 architectures however if you want to build OPA yourself, theMakefile
does not prohibit it.
- The diagnostics buffer in the OPA server has been completely removed as part of the deprecation and removal of the diagnostic feature (#1052)
- ast: Fix print call rewriting for calls in head (#3967)
- runtime: Fix logging configuration (#3959) (#3958)
This release includes a number of enhancements and fixes. In particular, this
release adds a new keyword for membership and iteration (in
) and a specialized
built-in function (print
) for debugging.
This release adds a new in
operator that provides syntactic sugar for
references that perform membership tests or iteration on collections (i.e.,
arrays, sets, and objects.) The following table shows common patterns for arrays
with the old and new syntax:
Pattern | Existing Syntax | New Syntax |
---|---|---|
Check if 7 exists in array | 7 == arr[_] |
7 in arr |
Check if 7 does not exist in array | n/a (requires helper rule) | not 7 in arr |
Iterate over the elements of array | x := arr[_] |
some x in arr |
For more information on the in
operator see Membership and iteration:
in
in the docs.
This release adds a new print
function for debugging purposes. The print
function can be used to output any value inside of the policy. The print
function has special handling for undefined values so that execution does not
stop if any of the operands are undefined. Instead, a special marker is emitted
in the output. For example:
package example
default allow = false
allow {
print("the subject's username is:", input.subject.username)
input.subject.username == "admin"
}
Given the policy above, we can see the output of the print
function via STDERR when using opa eval
:
echo '{"subject": {"username": "admin"}}' | opa eval -d policy.rego -I -f pretty 'data.example.allow'
Output:
the subject's username is: admin
true
If the username, subject, or entire input document was undefined, the print
function will still execute:
echo '{}' | opa eval -d policy.rego -I -f pretty 'data.example.allow'
Output:
the subject's username is: <undefined>
false
The print
function is integrated into the opa
subcommands, REPL, server, VS
Code extension, and the playground. Library users must opt-in to print
statements. For more information see the
Debugging
section in the docs.
- SDK: Allow map of plugins to be passed to SDK (#3826) authored by @edpaget
opa test
: Change exit status when tests are skipped (#3773) authored by @kirk-patton- Bundles: Improve loading performance (#3860) authored by @0xAP
opa fmt
: Keep new lines in between function arguments (#3836) reported by @anbrsapopa inspect
: Add experimental subcommand for bundle inspection (#3754)
-
Bundles/API: When deleting a policy, the check determining if it's bundle-owned was using the path prefix, which would yield false positives under certain circumstances. It now checks the path properly, piece-by-piece. (#3863 authored by @edpaget
-
CLI: Using
--set
with null value again translates to empty object (#3846) -
Rego: Forbid dynamic recursion with hidden (
system.*
) document (#3876 -
Rego: Raise conflict errors in functions when output not captured (#3912)
This change has the potential to break policies that previously evaluated successfully! See Backwards Compatibility notes below for details.
-
Experimental disk storage: React to "txn too big" errors (#3879), reported and authored by @floriangasc
- Kubernetes and Istio: Update tutorials for recent Kubernetes versions (#3910) authored by @olamiko
- Deployment: Add section about Capabilities (#3769)
- Built-in functions: Add warning to
http.send
and extension docs about side-effects in other systems (#3922) (#3893) - Docker Authorization: The tutorial now uses a Bundles API server.
- SDK: An example of SDK use is provided.
- Runtime: Refactor logger usage -- see below for Backwards Compatibility notes.
- Wasm: fix an issue with undefined, plain
input
references (#3891) - test/e2e: Extend TestRuntime to avoid global fixture
- types: Fix Arity function to return zero when type is known (#3932)
- Wasm/builder: bump LLVM to 13.0.0, latest versions of wabt and binaryen (#3908)
- Wasm: deal with importing memory in the compiler (#3763)
-
Function return values need to be well-defined: for a single input
x
, the function's outputf(x)
can only be one value. When evaluating policies, this condition had not been ensured for function calls that don't make use of their values, likepackage p r { f(1) } f(_) = true f(_) = false
Before,
data.p.r
evaluated totrue
. Now, it will (correctly) return an error:eval_conflict_error: functions must not produce multiple outputs for same inputs
In more realistic settings, this can be encountered when true/false return values are captured and returned where they don't need to be:
package p r { f("any", "baz") } f(path, _) = r { r := path == "any" } f(path, x) = r { r := glob.match(path, ["/"], x) }
In this example, any function input containing
"any"
would make the function yield two different results:- The first function body returns
true
, matching the"any"
argument. - The second function body returns the result of the
glob.match
call --false
.
The fix here would be to not capture the return value in the function bodies:
f(path, _) { path == "any" } f(path, x) { glob.match(path, ["/"], x) }
- The first function body returns
-
The
github.com/open-policy-agent/opa/runtime#NewLoggingHandler
function now requires a logger instance. Requiring the logger avoids the need for the logging handler to depend on the global logrus logger (which is useful for test purposes.) This change is unlikely to affect users.
This is a bugfix release addressing an issue in the formatting of rego code that contains object literals. With the last release, those objects would under some conditions have their keys re-ordered, with some of them put into a single line.
Thanks to @iainmcgin for reporting.
- format: make groupIterable sort by row (#3849)
This release includes a number of improvements and fixes.
This release introduces crypto.x509.parse_rsa_private_key
so that policy authors can decode RSA private keys and structure them as JWKs (#3765). Authored by @cris-he.
- Fix object comparison to avoid sorting keys in-place. This prevents the interpreter from generating non-deterministic results when values are inserted into the partial set memoization cache. (#3819)
- Fix data races in
ast
package caused by sortingtypes.Any
instances in-place and shallow-copying module comments when a deep-copy should be performed (#3793). Reported by @markushinz. - Fix "file name too long" error caused by bundle loader treating PEM encoded private keys as file paths (#3766)
- Fix plugins to support manual triggering mode when discovery is disabled (#3797)
- The server now supports policy-based health checks that can inspect the state of plugins and other internal components (#3759) authored by @gshively11
- The bundle reader now loads files lazily to avoid hitting file descriptor limits (#3777). Authored by @bhoriuchi
- The
opa eval
sub-command supports a--timeout
option for limiting how long evaluation can run.
- The type checker now supports variadic arguments on void functions. This change paves the way for
print()
support as well as variadic arguments on all functions. - The parser now memoizes term parsing. This prevents non-linear runtime for large nested objects and sets.
- Fix spurious build errors in wasm library.
- Update wasmtime dependency to v0.30.0.
- Run PR checks on macOS in addition to Linux (#3176).
- Update the Kubernetes and Envoy (standalone) tutorials to show how the OPA management APIs can be used to distribute policies.
- The
github.com/open-policy-agent/opa/ast#ArgErrDetail
struct has been modified to use the newtypes.FuncArgs
struct to represent the required arguments. Callers that depend on the exact structure of the error details must update to use thetypes.FuncArgs
struct.
This is a bugfix release to address a problem related to mismatching checksums in the official go mod proxy. As a consequence, users with code depending on the OPA Go module that bypassed the proxy would see an error like
go get github.com/google/flatbuffers/go: github.com/google/flatbuffers@v1.12.0: verifying module: checksum mismatch
downloaded: h1:N8EguYFm2wwdpoNcpchQY0tPs85vOJkboFb2dPxmixo=
sum.golang.org: h1:/PtAHvnBY4Kqnx/xCQ3OIV9uYcSFGScBsWI3Oogeh6w=
Be aware that Github's Dependabot feature makes use of that check, and will start to fail for projects using the OPA Go module version 0.32.0.
There workaround applied to OPA is to replace to flatbuffers dependency's version manually.
For more information, see
- google/flatbuffers#6466: The issue has been discussed upstream, and a 1.12.1 release has been published to address it.
- dgraph-io/badger#1746: OPA transitively depends on the flatbuffer package because of badger.
There are no functional changes in this bugfix release. If you use the container images, or the published binaries, of OPA 0.32.0, you are not affected by this.
Many thanks to James Alseth for triaging this, and engaging with upstream to fix this.
This release includes a number of improvements and fixes.
This release adds a disk-based storage implementation to OPA. The implementation can be found in github.com/open-policy-agent/storage/disk. There is also an example in the rego
package that shows how policies can be evaluated with the disk-based store. The disk-based store is currently only available as a library (i.e., it is not integrated into the rest of OPA yet.) In the next few releases, we are planning to integrate the implementation into the OPA server and provide tooling to help leverage the disk-based store.
This release includes a few improvements to existing built-in functions:
- The
http.send
function now supports UNIX domain sockets (#3661) authored by @kirk-patton - The
units.parse_bytes
function now supports E* and P* units (#2911) - The
io.jwt.encode_sign
function uses the built-in context randomization source (which is helpful for replay purposes)
This release includes multiple improvements for OPA server deployments in serverless environments:
- Plugins can now be triggered manually within OPA. This feature allows users extending and customizing OPA to control exactly when operations like bundle downloads and decision log uploads occur. The built-in plugins now include a
trigger
configuration that can be set tomanual
orperiodic
(which is the default). Whenmanual
triggering is enabled, the plugins WILL NOT perform any periodic/background operations. Instead, the plugins will only execute when theTrigger
API is invoked. - Plugins can now wait for server initialization. When runtime initialization is finished, plugins can be notified. This allows plugins to synchronize their behaviour with server startup. #3701 authored by @gshively11.
- The Health API now supports an
exclude-plugin
parameter to control which plugins are checked. #3713 authored by @gshively11.
- The compiler no longer fetches remote schemas by default when used as as library. Capabilities have been updated to include an
allow_net
field to control whether network operations can be performed (#3746). This field is only used to control schema fetching today. In future versions of OPA, theallow_net
parameter will be used to control other behaviour likehttp.send
. - The
WebAssembly runtime not supported
error message has been improved #3739.
- Added support for
anyOf
andallOf
keywords in JSON schema support in the type checker (#3592) authored by @jchen10500 and @juliafriedman8. - Added support for custom JSON result marshalling in the
rego
package. - Added a new convenience function (
Allowed() bool
) to therego.ResultSet
API. - Improved string-representation construction performance for arrays, sets, and objects.
- Improved the topdown evaluator to support
ast.Value
results from the store so that unnecessary conversions can be avoided. - Improved the
rego
package to make the wasmtime-go dependency optional at build-time (#3545). - Fixed a bug in the comprehension indexer whereby index keys were not constructed correctly leading to incorrect outputs (#3579).
- Fixed a stack overflow during partial evaluation due to incorrect term rewriting in the copy propagation implementation (#3071).
- Fixed a bug in partial evaluation when shallow inlinign is enabled that resulted in built-in functions being invoked instead of saved (#3681).
- The internal Wasm SDK now supports the inter-query built-in cache.
- The pre-compiled runtime is now built with llvm 12.0.1 and the builder image includes clang-format.
- The internal Wasm SDK has been updated to use wasmtime-go v0.29.0.
This release includes a number of documentation improvements:
- The wasm
opa_eval
arguments have been clarified #3699 - The contributing and development guide have been moved into a dedicated Contributing section on the website #3751
- The Envoy standalone tutorial includes cleanup steps now (thanks @princespaghetti)
- Various typos have been fixed by multiple folks (thanks @Tej-Singh-Rana @gujun4990)
- The Kubernetes ingress validation tutorial has been updated to include new mandatory attributes and newer API versions (thanks @ereslibre)
- The recommendations around using OPA Gatekeeper have been improved.
- OPA is now built with Go v1.17 and CI jobs have been added to ensure OPA builds with older versions of Go.
The rego
package no longer relies on build constraints to enable the Wasm runtime. Instead, library users must opt-in to Wasm runtime support by adding an import statement in the Go code:
import _ "github.com/open-policy-agent/opa/features/wasm"
This change ensures that (by default) the wasmtime-go blobs are not vendored in projects that embed OPA as a library. If you are currently relying on the Wasm runtime support in the rego
package (via the rego.Target("wasm")
option), please update you code to include the import above. See #3545 for more details.
This release contains performance improvements for evaluating partial sets and objects, and introduces a new ABI call to OPA's Wasm modules to speed up Wasm evaluations.
It also comes with an improvement for checking policies -- unsafe declared variables are now caught at compile time. This means that some policies that have been working fine with previous versions, because their unsafe variables had not ever been queried, will fail to compile with OPA 0.31.0. See below for details and what to do about that.
Resolving an issue (#822) created on July 4th 2018, OPA can now cache the results of partial sets and partial objects.
A benchmark that accesses a partial set of increasing size twice shows a saving of more than 50%:
name old time/op new time/op delta
PartialRuleSetIteration/10-16 230µs ±10% 101µs ± 3% -56.10% (p=0.000 n=10+10)
PartialRuleSetIteration/100-16 13.4ms ± 9% 5.5ms ± 9% -58.74% (p=0.000 n=10+9)
PartialRuleSetIteration/1000-16 1.31s ±10% 0.51s ± 8% -61.12% (p=0.000 n=10+9)
name old alloc/op new alloc/op delta
PartialRuleSetIteration/10-16 77.7kB ± 0% 35.8kB ± 0% -53.94% (p=0.000 n=10+10)
PartialRuleSetIteration/100-16 3.72MB ± 0% 1.29MB ± 0% -65.26% (p=0.000 n=10+10)
PartialRuleSetIteration/1000-16 365MB ± 0% 114MB ± 0% -68.86% (p=0.000 n=10+10)
name old allocs/op new allocs/op delta
PartialRuleSetIteration/10-16 1.84k ± 0% 0.69k ± 0% -62.42% (p=0.000 n=10+10)
PartialRuleSetIteration/100-16 99.3k ± 0% 14.5k ± 0% -85.43% (p=0.000 n=10+9)
PartialRuleSetIteration/1000-16 10.0M ± 0% 1.0M ± 0% -89.58% (p=0.000 n=10+9)
These numbers were gathered querying fixture[i]; fixture[j]
with a policy of
fixture[x] {
x := numbers.range(1, n)[_]
}
where n
is 10, 100, or 1000.
There are multiple access patterns that are accounted for: if a ground scalar is used to access a previously not-cached partial rule,
allow {
managers[input.user] # here
}
managers[x] {
# some logic here
}
the evaluation algorithm will calculate the set membership of input.user
only, and cache the result.
If there is a query that requires evaluating the entire partial, however, the algorithm will also cache the entire partial:
allow {
some person
managers[person]
# more expressions
}
managers[x] {
# some logic here
}
thus avoiding extra evaluations later on.
The same is true if managers
was used as a fully materialized set in an execution.
This also means that the question about whether to write
q = { x | ... } # set comprehension
or
q[x] { ... } # partial set rule
becomes much less important for policy evaluation performance.
OPA-generated Wasm modules have gotten a fast-path evaluation method: By calling the one-off function
opa_eval(reserved, entrypoint, data_addr, input_addr, input_len, format)
which returns a pointer to the serialized result set (in JSON if format is 0, "value" format if 1), the number of VM calls needed for evaluating a policy via Wasm is drastically reduced.
The performance benefit is huge:
name old time/op new time/op delta
WasmRego-16 84.3µs ± 6% 15.1µs ± 0% -82.07% (p=0.008 n=5+5)
The added opa_eval
export comes with an ABI bump to version 1.2.
See #3627 for all details.
Along the same line, we've examined the processing of query evaluations that are Wasm-backed through the rego
package.
This allowed us to avoid unneccessary work (#3666).
Before this release, local variables that had been declared, i.e. introduced via the some
keyword, had been able
to slip through the safety checks unnoticed.
For example, a policy like
package demo
q {
input == "open sesame"
}
p[x] {
some x
}
would have not caused any error if data.demo.p
wasn't queried.
Querying data.demo.p
would return an "var requires evaluation" error.
With this release, the erroneous rule no longer goes unnoticed, but is caught at compile time: "var x is unsafe".
The most likely fix is to remove the rule with the unsafe variable, since it cannot have contributed to a successful evaluation in previous OPA versions.
See #3580 for details.
-
New built-in function:
crypto.x509.parse_and_verify_certificates
(#3601), authored by @jalsethThis function enables you to verify that there is a chain from a leaf certificate back to the trusted root.
-
New built-in function:
rand.intn
generates a random number between0
andn
(#3615), authored by @base698The function takes a string argument to ensure that the same call, within one policy evaluation, returns the same random number.
-
http.send
enhancement: Newcaching_mode
parameter to configure if deserialized or serialized response bodies should be cached (#3599) -
Custom built-in function enhancement: let custom builtins halt evaluation (#3534)
-
Partial evaluation: Fix stack overflow on certain expressions (#3559)
-
Query Profiling:
opa eval --profile
now supports a--count=#
flag to gather metrics and profiling data over multiple runs, and displays aggregate statistics for the results (#3651).This allows you to gather more robust numbers to assess policy performance.
-
Docker images: Publish static image (#3633)
As of this release, you can use the staticly-built Linux binary from a docker image:
openpolicyagent/opa:0.31.0-static
. It contains the same binary that has been published since release v0.29.4, statically linked to musl, with evaluating Wasm disabled.
- Built-in
http.send
: ignoretls_use_system_certs
setting on Windows. Having this set to true (the default as of v0.29.0) would always return an error on Windows. - The console decision logger is no longer tied to the general log level (#3654)
- Update query compiler to reject empty queries (#3625)
- Partial Evaluation fix: Don't generate comprehension with unsafe variables (#3557)
- Parser: modules containing only tabs and spaces no longer lead to a runtime panic.
- Wasm: ensure that the desired stack space for the C library calls (64KiB) is not reduced by data segments added in the compiler. This is achieved by putting the stack first -- stack overflows now become "out of bounds" memory access traps. Before, it would silently corrupt the static data.
-
New configuration for Management APIs: using
resource
, the request path for sending decision logs can be configured now (#3618), authored by @cbuto/logs
is still the default, but can now be overridden. With this change, thepartition_name
config becomes deprecated, since its functionality is subsumed by this new configurable.
- How to debug? Clarify how to access
Note
events for debugging via explanations (#3628) authored by @enori - Clarify special characters for key, i.e. what
x["y"]
is necessary becausex.y
isn't valid (#3638) authored by @Hongbo-Miao - Management APIs: Remove deprecated fields from docs
- Policy Reference: add missing backtick;
type_name
builtin is natively implemented in Wasm
This is a bugfix release that modifies the AWS credential provider to use POST instead of GET for retrieving AWS STS tokens. The GET method can leak credentials into the debug log if the AWS STS endpoint is unavailable.
This is a bugfix release to correct the behaviour of the indexof
builtin (#3606).
In v0.30.0, it only checked the first character of the substring to be found: indexof("foo", "fox")
erroneously returned 0 instead of -1.
- wasm-sdk: Fix typo in non-wasm error message, authored by @olivierlemasle
This release contains a number of enhancements and fixes.
-
Support listening on abstract Unix Domain Sockets (#3533) authored by @amanymous-net
-
Support minimum TLS version configuration, default to 1.2 (#3226) authored by @kale-amruta
-
Enhancement in REST Plugin: You can now specify a CA cert for remote services implementing the management APIs (bundles, status, decision logs, discovery) (#1954)
-
Bugfix: treat missing/empty roots as owning all paths (#3521)
Before, it would have been possible to overwrite a policy that was supplied by a bundle (with an empty manifest, or a manifest without declared roots), due to an erroneous check. This will now be forbidden, and return a 400 HTTP status, in accordance with the documentation.
-
Extend POST v1/query endpoint to accept input, refactor index.html to use fetch()
-
Bundle download: In case of download or activation errors, the cached Etag is reset to the last successful activation. Previously OPA would reset the cached Etag entirely, which could trigger unnecessary bundle downloads in edge-case scenarios.
opa build
: Do not write manifest if empty (#3480). Under the hood, the manifest metadata is now included in the Equal() function's checks.opa fmt
: Fix incorrect help text (#3518) authored by @andrehalandopa bench
: Do not print nil errors (#3530)
- Expose random seeding in rego package (#3560)
- Enhance
ast.InterfaceToValue
to handle non-native types - Enhance indexer to understand function args
- Enhance static property lookup of objects: Use binary search
- Fix PE unknown check to avoid saving unnecessarily (#3552)
- Fix inlining controls for functions (#3463)
- Fix (shallow) partial eval of ref to empty collection in presence of
with
statement (#3420) - Fix cache value size checking during insert operation
- Fix
indexof
when using UTF-8 characters - Fix
http.send
flaky test
- SDK: update wasmtime-go to 0.28.0, authored by @olivierlemasle
- Bugfix: count() now counts invalid UTF-8 runes (previously aborted)
- Compiler: emit unreachable instruction after opa_abort()
make check
now uses golangci-lint via docker, authored by @willbeason- The statically-built linux binary is properly used in the make targets that need it, and published to edge binaries.
- Built binaries are now smoke tested on Windows, macos, and Linux.
- Fix test failing with Go 1.17 rc in gojsonschema (#3589) authored by @olivierlemasle
- Build: Bump Go version to 1.16.3 (#3555)
- CI: enable dependabot for wasmtime-go
- OAuth2/OIDC: Fixed
concat
arguments in metadata discovery method (#3543, @iggbom) - Policy Reference: syntax highlighting EBNF grammar (@PatMyron)
- Extending OPA: fix typo (@dxps)
- Extending OPA: marshal the decision log (@TheLunaticScripter)
- Kubernetes Introduction: fix typo (@dbaker-rh)
- Envoy: Add guidance for OPA-Envoy benchmarks
- Change default linux download to
opa_linux_amd64_static
This is a bugfix release that re-introduces linux binaries that do not depend on glibc, i.e., run in unmodified Alpine Linux systems.
- bundle: Implement a DirectoryLoader for fs.FS (#3493) (#3489) authored by @simongottschlag
This bugfix release addresses another edge case in function evaluation (#3505).
This is a bugfix release to resolve an issue in topdown's function output caching (#3501)
This is a bugfix release to resolve an issue in the release pipeline.
This release contains a number of enhancements and fixes.
- This release includes a new top-level package to support OPA integrations in Go programs:
github.com/open-policy-agent/opa/sdk
. Users that want to integrate OPA as a library in Go and expose features like bundles and decision logging should use this package. The package is controlled by specifying an OPA configuration file. Hot reloading is supported out-of-the-box. See the GoDoc for the package docs for more details.
- A deadlock in the bundle plugin during shutdown has been resolved (#3363)
- An issue between bundle signing and bundle persistence when multiple data.json files are included in the bundle has been resolved (#3472)
- The
github.com/open-policy-agent/opa/runtime#Params
struct now supports a router parameter to enable custom routes on the HTTP server. - The bundle manifest can now include an extra
metadata
key where arbitrary key-value pairs can be stored. Authored by @viovanov - The bundle plugin now supports file:// urls in the
resource
field for test purposes. - The decision log plugin emits a clearer message at DEBUG instead of INFO when there is no work to do. Authored by andrewbanchich
- The discovery plugin now supports a
resource
configuration field like the bundle plugin. Similarly, theresource
is treated as the canonical setting to identify the discovery bundle.
- The
opa test
timeout as been increased to 30 seconds when benchmarking (#3107) - The
opa eval --schema
flag has been fixed to correctly set the schema when a single schema file is passed - The
opa build --debug
flag output has been improved for readability - The
array.items
JSON schema value is now supported by the type checker - The
opa fmt
subcommand can now exit with a non-zero status when a diff is detected (by passing--fail
) - The
opa test
subcommand no longer emits bogus file paths when fed a file:// url
- The
http.send
built-in function falls back to the system certificate pool when thetls_ca_cert
ortls_ca_cert_env_variable
options are not specified (#2271) authored by @olamiko
- The order of support rules emitted by partial evaluation is now deterministic (#3453) authored by @andrehaland
- The big number performance regression caught by the fuzzer has been resolved (#3262)
- The evaluator has been updated to memoize calls to rules with arguments (functions) within a single query. This avoids recomputing function results when the same input is passed multiple times (similar to how complete rules are memoized.)
- The
wasm
target no longer panics if the OPA binary does not include a wasm runtime (#3264) - The interrupt handling mechanism has been rewritten to make safe use of the wasmtime package. The SDK also returns structured errors now that are more aligned with topdown. (#3225)
- The SDK provides the subset of required imports now (which is useful for debugging with opa_println in the runtime library if needed.)
- The opa_number_float type has been removed from the value library (it was unused after moving to libmpdec)
- The runtime library builder has been updated to use llvm-12 and the wasmtime-go package has been updated to v0.27.0
- The HTTP API authorization tutorial has been updated to show how to distribute policies using bundles
- The Envoy tutorial has been tweaked to show better path matching examples
- The release-patch script has been improved to deal with this file in bugfix/patch releases (#2533) authored by @jjshanks
- The Makefile check targets now rely on golangci-lint and many linting errors have been resolved (authored by @willbeason)
- Multiple nightly fuzzing and data race issues in test cases have been resolved
This release includes a number of features, enhancements, and fixes. The default
branch for the Git repository has also been updated to main
.
This release adds support for annotations. Annotations allow users to declare metadata on rules and packages. Currently, OPA supports one form of metadata: schema declarations. For example:
package example
# METADATA
# schemas:
# - input: schema.service
deny["service is missing required 'owner' label"] {
input.kind == "Service"
not input.metadata.labels.owner
}
# METADATA
# schemas:
# - input: schema.deployment
deny["deployment replica count too low for 'production' namespace"] {
input.kind == "Deployment"
input.metadata.namespace == "production"
object.get(input.spec, "replicas", 1) < 3
}
Users can include schema annotations in their policies to tell OPA about the
structure of external data loaded under input
or data
. By learning the
schema of base documents, OPA can surface mistakes in the policy at authoring
time (e.g., referring to a non-existent field in a JSON object or calling a
built-in function with an invalid value.) For more information on the
annotations and schema support see the Type
Checking page in the
documentation. In the future, annotations will be expanded to support other
kinds of metadata and additional tooling will be added to leverage them.
- The server now automatically sets GOMAXPROCS when running inside of a container that has cgroups applied. This helps the Go runtime avoid consuming too many CPU resources and being throttled by the kernel. (#3328)
- The server now logs an error if users enable the
token
authentication mode without a corresponding authorization policy. (#3380) authored by @kale-amruta - The server now supports a
GET /v1/config
endpoint that returns OPA's active configuration. This API is useful if you need to debug the running configuration in an OPA configured via Discovery. (#2020) - The server now respects the
?pretty
option in the v0 API (#3332) authored by @clarshad - The Bundle plugin is more forgiving when it comes to Etag processing on HTTP 304 responses (#3361)
- The Decision Log plugin now supports a "Decision Per Second" rate limit configuration setting.
- The Status plugin can now be configured to use a custom reporter similar to the Decision Log plugin (e.g., so that Status messages can be sent to AWS Kinesis, etc.)
- The Status plugin now reports the number of decision logs that are dropped due to buffer limits.
- The service clients can authenticate with the Azure Identity OAuth2 implementation the client credentials JWT flow is used (#3372)
- Library users can now customize the logger used by the plugins by providing the
plugins.Logger
option when creating the plugin manager.
- The various OPA subcommands that accept schema files now accept a directory tree of schemas instead of only a single schema.
- The
opa refactor move
subcommand was added to support package renaming use cases (#3290) - The
opa check
subcommand now supports a-s
/--schema
flag like theopa eval
subcommand.
- The Management API docs have been restructured so that each API has a dedicated page. In addition, the Bundle API docs now include getting started steps for cloud-provider specific services (e.g., AWS, GCP, Azure, etc.)
- OPA now supports PKCS8 encoded EC private keys for JWT verification (which includes service authentication, bundle verification, and verification built-in functions) (#3283). Authored by @andrehaland.
- The bundle signing and verification APIs have been updated to support custom signers/verififers (#3336). Authored by @gshively11.
- The
time.diff
function was added to support calculating differences between date/time values (#3348) authored by @andrehaland - The
units.parse_bytes
function now supports floating-point values (#3297) authored by @andy-paine - The evaluator was fixed to use correct bindings when evaluating the full-extent of a partial rule set. This issue was causing unexpected undefined results and evaluation errors in some rare cases. (#3369 #3376)
- The evaluator was fixed to correctly generate package paths when namespacing is disabled partial evaluation. (#3302).
- The
http.send
function no longer errors out on invalid Expires headers. (#3284) - The inter-query cache now serializes elements on insertion thereby reducing memory usage significantly (because deserialized elements carry a ~20x cost.) (#3042)
- The rule indexer was fixed to correctly handle mapped and non-mapped values which could occur with
glob.match
usage (#3293)
- The
opa eval
subcommand now correctly returns the set of all variable bindings and expression values when thewasm
target is enabled. Previously it returned only set of variable bindings. (#3281) - The
glob.match
function now handles the default delimiter correctly. (#3294) - The
opa build
subcommand no longer requires a capabilities file when thewasm
target is enabled. If capabilities are not provided, OPA will use the capabilities for its own version. (#3270) - The
opa build
subcommand now dumps the IR emitted by the planner when--debug
is specified. - The
opa eval
subcommand no longer panics when a policy fails to type check and thewasm
target is enabled. - The comparison functions can now return
false
instead of either beingtrue
orundefined
. (#3271) - The internal wasm runtime will now correctly return
CancelErr
to indicate cancellation errors (instead ofBuiltinErr
which it returned previously.) - The internal wasm runtime now correctly handles non-halt built-in errors (#3320)
- The planner no longer generates unexpected scan statements when negation used over base documents under
data
(#3279) and (#3305) - The planner now correctly discards out-of-scope variables when exiting comprehensions (#3325)
- The
rego
package no longer panics when thewasm
target is enabled and undefined functions are encountered (#3251) - 🎈 The remaining exceptions in the e2e test framework for the internal wasm runtime have been resolved.
- The
make image
target now uses the CI image for building the Go binary. This avoids platform-specific build issues by building the Go binary inside of Docker.
This release contains a fix for crashes experienced when configuring OPA to use S3 signing as service credentials (#3255).
In addition to that, we have a small number of enhancements and fixes:
- The
eval
subcommand now allows using--import
without using--package
. Authored by @onelittlenightmusic, #3240.
- The
ast
package now exports another method for JSON conversion,ast.JSONWithOpts
, that allows further options to be set (#3244.
- REST plugins using
s3_signing
as credentials method can now include the specified service in the signature (SigV4). Authored by @cogwirrel, #3210.
- Remove soon-to-be deprecated
any
andall
from the Policy Reference (#3241) -- see also #2437. - Add missing
discovery.service
field to Discovery configuration table (#3237). - Fix dead links to the Envoy pages (#3248).
- Executions using the internal Wasm SDK will now be interrupted when the provided context is done (cancelled or deadline reached).
- The generated Wasm modules could become much smaller: unused functions are replaced by
unreachable
stubs, and the heavyweight runtime components related to regular expressions are excluded when none of the regex-related builtins are used:glob.match
,regex.is_valid
,regex.match
,regex.is_valid
, andregex.find_all_string_submatch_n
. - The Wasm runtime now allows passing in the time to be used for evaluation, enabling callers to control the time-of-day observed by Wasm compiled policies.
- Wasmtime runtime has been updated to the latest version (v0.24.0).
This release contains a number of enhancements and bug fixes.
- The
eval
subcommand now supports a-s
/--schema
flag that accepts a JSON schema for theinput
document. The schema is used when type checking the policy so that invalid references to (or operations on)input
data are caught at compile time. In the future, the schema support will be expanded to accept multiple schemas and rule-level annotations. See the new Schemas documentation for details. Authored by @aavarghese and @vazirim. - The
eval
,test
,bench
and REPL subcommands now supports a-t
/--target
flag to set the evaluation engine to use. The default engine isrego
referring to the standard Rego interpreter in OPA. Users can now selectwasm
to enable Wasm compilation and execution of policies (#2878). - The
eval
subcommand now supports araw
option for-f
/--format
that is useful in bash scripts. Authored by @jaspervdj-luminal. - The test framework now supports "skippable" tests. Prefix the test name with
todo_
to have the test runner skip the test, e.g.,todo_test_allow { ... }
. - The
eval
subcommand now correctly supports the--ignore
flag. Previously the flag was not being applied.
- The
POST /v1/compile
API now supports a?metrics
query parameter similar to other APIs. Authored by @jkbschmid. - The directory used for persisting downloaded bundles can now be configured. See the Configuration page for details.
- The HTTP Decision Logger plugin no longer blocks server shutdown for the grace period when there are no logs to upload.
- The Bundle plugin now unregisters listeners correctly. This issue would cause listeners to be invoked when bundle updates were dispatched even if the listener was unregistered (#3190).
- The server now correctly decodes policy IDs in the HTTP request URL. Authored by @mattmahn (#2116).
- The server now configures the
http_request_duration_seconds
metric (for all of the server endpoitns) with smaller, more granular buckets that better map to actual response latencies from OPA. Authored by @luong-komorebi (#3196).
- PKCS8 keys are now supported when signing bundles and communicating with control plane services. Previously only PKCS1 keys were supported (#3116).
- The built-in OPA HTTP API authorizer policy can now return a reason to explain why a request to the OPA API is denied (#3056). See the Security documentation for details. Thanks to @ajanthan for helping improve this.
- The compiler can be configured to emit debug messages that explain comprehension indexing decisions. Debug messages can be enabled when running
opa build
with--debug
. - A panic was fixed in one of the rewriting stages when comprehensions were used as object keys (#2915)
- A bug in big integer comparison was fixed. This issue was discovered when comparing serial numbers from X.509 certificates. Authored by @andrehaland (#3147).
- The
io.jwt.decode_verify
function now uses the environment supplied time-of-day value instead of callingtime.Now()
(#3105).
- The documentation now includes a dedicated section the OPA-Envoy integration. See https://www.openpolicyagent.org/docs/latest/envoy-introduction/ for details.
- The ecosystem page now ranks integrations by number of unique domains instead of the sheer number of references.
- The
data
document no longer needs to be initialized to an empty object (#3130). - The mpd library is now initalized by the module's
Start
function (#3110). - The planner now longer re-plans rules blindly when
with
statements are encountered (#3150). - The planner and compiler now support dynamic dispatch. Previously the planner would enumerate all functions and invocation was controlled at runtime (#2936).
- The compiler now inserts memoization instructions into function bodies instead of at callsites. This reduces the number of wasm instructions in the resulting binary (#3169).
- The wasmtime runtime is now the default runtime used by OPA to execute compiled policies. The new runtime no longer leaks memory when policies are reloaded.
- The planner and compiler now intern strings and booleans and implement a few micro-optimizations to reduce the size of the resulting binary.
- The capabilities support has been updated to include an ABI major and minor version for tracking backwards compatibility on compiled policies (#3120).
- The
opa test
subcommand previously supported a-t
flag as shorthand for--timeout
. With this release, the-t
shorthand has been redefined for--target
. After searching GitHub for examples ofopa test -t
(and finding nothing) we felt comfortable making this backwards incompatible change. - The Go version used to build the OPA release has been updated from
1.14.9
to1.15.8
. Because of this, TLS certificates that rely on Common Name for verification are no longer supported and will not work. For more information see golang/go#39568.
This release contains a number of enhancements and bug fixes.
-
This release includes a number of built-in function improvements for Wasm compiled policies. The following built-in functions have been implemented natively and no longer need to be supplied by SDKs:
graph.reachable
,json.filter
,json.remove
,object.get
,object.remove
, andobject.union
. -
This release fixes several bugs in the Wasm implementation of certain
regex
built-in functions (#2962),format_int
(#2923) andround
(#2999). -
This release adds
ceil
andfloor
built-in functions. Previously these could be implemented in Rego usinground
however these are more convenient.
-
OPA has been extended support OAuth2 JWT Bearer Grant Type and OAuth2 Client Credential JWT authentication options for communicating with control plane services. This change allows OPA to use services that rely on Ping Identity as well as GCP service accounts for authentication. OPA has also been extended to support custom authentication plugins (thanks @gshively11).
-
OPA plugins can now enter a "WARN" state to indicate they are operating in a degraded capacity (thanks @gshively11).
-
The
opa bench
command can now benchmark partial evaluation queries. The options to enable partial evaluation are shared withopa eval
. Seeopa bench --help
for details. -
Wasm compiled policies now contain source locations that are included inside of runtime error messages (such as object key conflicts.) In addition, Wasm compiled policies only export the minimal set of APIs described on the WebAssembly#exports page.
- ast: Fix parsing of numbers to reject leading zeroes (#2947) authored by @LCartwright.
- bundle: Fix loader to only verify bundle keys if configured to do so (#3028).
- cmd: Fix build to avoid packaging policy.wasm twice (#3007).
- cmd: Fix pretty-printed PE output to hide spurious blank lines
- server: Fix false-positive in bundle root check that would prevent data updates in some cases (#2868).
- server: Fix query cache to respect ?instrument option (#3000).
- server: Fix server to support discovery on inter-query cache configuration
- topdown: Fix PE to avoid generating expressions that do not type check (#3012).
- wasm: Fix planner to avoid generating a conflict error in some cases (#2926).
- wasm: Fix planner to generate correct virtual document iteration instructions (#3065).
- wasm, topdown: Fix with keyword handle to ensure last statement wins (#3010).
- wasm: Fix planner to handle assignment conflicts correctly when else keyword is used (#3031).
- Add new section on integrating policies with OAuth2 and OIDC.
- Update Kubernetes admission control tutorial to work as non-root user.
- Fix link to signing documentation (#3027) authored by @princespaghetti.
- Previously, OPA deduplicated sets and objects in all cases except when iterating over/referring directly to values generated by partial rules. This inconsistency would only be noticed when running ad-hoc queries or within policies when aggregating the results of array comprehensions (e.g.,
count([1 | p[x]])
could observe duplicates inp
.) This release removes the inconsistency by deduplicating sets and objects in all cases (#429). This was the second oldest open issue on the project.
- OPA now logs warnings when it receives legacy
bundle
config sections instead of thebundles
section introduced in v0.13.0.
This release extends the HTTP server authorizer (--authorization=basic
) to supply the HTTP message body in the input
document. See the Authentication and Authorization section in the security documentation for details.
This release contains a fix for running OPA under Docker with a non-default working directory (#2974).
This release contains a number of improvements and fixes. Importantly, this release includes a notable change to built-in function error handling. See the section below for details.
Previously, built-in function errors would cause policy evaluation to halt immediately. Going forward, by default, built-in function errors no longer halt evaluation. Instead, expressions are treated as false/undefined if any of the invoked built-in functions return errors.
This change resolves a common issue people face when passing unsanitized input values to built-in functions. For example, prior to this change the expression io.jwt.decode("GARBAGE")
would halt evaluation of the entire policy because the string is not a valid encoding of a JSON Web Token (JWT). If the expression was io.jwt.decode(input.token)
and the user passed an invalid string value for input.token
the same error would occur. With this change, the same expression is simply undefined, i.e., there is no result. This means policies can use negation to test for invalid values. For example:
decision := {"allowed": allow, "denial_reason": reason}
default allow = false
allow {
io.jwt.verify_hs256(input.token, "secret")
[_, payload, _] := io.jwt.decode(input.token)
payload.role == "admin"
}
reason["invalid JWT supplied as input"] {
not io.jwt.decode(input.token)
}
If you require the old behaviour, enable "strict" built-in errors on the query:
Caller | Example |
---|---|
HTTP | POST /v1/data/example/allow?strict-builtin-errors |
Go (Library) | rego.New(rego.Query("data.example.allow"), rego.StrictBuiltinErrors(true)) |
CLI | opa eval --strict-builtin-errors 'data.example.allow' |
If you have implemented custom built-in functions and require policy evaluation to halt on error in those built-in functions, modify your built-in functions to return the topdown.Halt error type.
This release includes a few new built-in functions:
base64url.encode_no_pad
,hex.encode
, andhex.decode
for dealing with encoded data (#2849) authored by @johanneslarssonjson.patch
for applying JSON patches to values inside of policies (#2839) authored by @jaspervdj-luminaljson.is_valid
andyaml.is_valid
for testing validity of encoded values (authored by @jaspervdj-luminal)
There were also a few fixes to existing built-in functions:
- Fix unicode handling in a few string-related functions (#2799) authored by @anderseknert
- Fix
http.send
to overrideno-cache
HTTP header whenforce_cache
specified (#2841) authored by @anderseknert - Fix
strings.replace_n
to replace overlapping patterns deterministically (#2822) - Fix panic in
units.parse_bytes
when passed a zero-length string (#2901)
This release adds new credential providers for management services:
- GCP metadata server (#2938) authored by @kelseyhightower
- AWS Web Identity credentials (#2462) authored by @RichiCoder1
- OAuth2 (#1205) authored by @anderseknert
In addition the following server features were added:
- Add shutdown wait period flag to
opa run
(--shutdown-wait-period
) (#2764) authored by @bcarlsson - Add bundle file size limit configuration option (
bundles[_].size_limit_bytes
) to override default 1GiB limit (#2781) - Separate decision log and status message logs from access logs (which useful for running OPA at log level
error
while continuing to report decision and status log to console) (#2733) authored by @anderseknert
- Fix panic caused by race condition in the decision logger (#2835) authored by @kubaj
- Fix decision logger to flush on graceful shutdown (#780) authored by @anderseknert
- Fix
--verification-key
handling to accept PEM files (#2796) - Fix
--capabilities
flag inopa build
command (#2848) authored by @srenatus - Fix loading of signed persisted bundles (#2824)
- Fix API response mutation caused by decision log masking (#2752) authored by @gshively11
- Fix evaluator to prevent
with
statements from mutating originalinput
document (#2813) - Fix set iteration runtime to be O(n) instead of O(n^2) (#2966)
- Increased OPA version telemetry report timeout from 1 second to 5 seconds to deal with slow networks
- Improve docs to mention built-in function support in WebAssembly compiled policies
- Improve docs around JWT HMAC encoding (#2870) authored by @anderseknert
- Improve HTTP authorization tutorial steps for zsh (#2917 authored by @ClaudenirFreitas)
- Improve docs to describe meaning of Prometheus metrics
- Remove mention of unsafe (and unsupported) "none" signature algorithm from JWT documentation
This release also includes a number of improvements to the Wasm support in OPA. Importantly, OPA now integrates a Wasm runtime that can be used to execute Wasm compiled policies. The runtime is integrated into the existing "topdown" evaluator so that specific portions of the policy can be compiled to Wasm as a performance optimization. When the evaluator executes a policy using the Wasm runtime it emits a special Wasm
trace event. The Wasm runtime support in OPA is currently considered experimental and will be iterated on in coming releases.
This release also extends the Wasm compiler in OPA to natively support the following built-in functions (in alphabetical order):
base64.encode
,base64.decode
,base64url.encode
, andbase64url.decode
glob.match
json.marshal
andjson.unmarshal
net.cidr_contains
,net.cidr_intersects
, andnet.cidr_overlap
regex.match
,regex.is_valid
, andregex.find_all_string_submatch_n
to_number
walk
- The
--insecure-addr
flag (which was deprecated in v0.10.0) has been removed completely (#763)
This release contains a number of small enhancements and bug fixes.
This release adds support for persisting bundles for recovery purposes. When persistence is enabled, OPA will save activated bundles to disk. On startup, OPA checks for persisted bundles and activates them immediately. This allows OPA to startup if the bundle server is unavailable (#2097). For more information see the Bundle documentation.
This release includes a few new built-in functions:
base64.is_valid
for testing if strings are valid base64 encodings (#2690) authored by @carlpettnet.cidr_merge function
for merging sets of IPs and CIDRs (#2692)urlquery.decode_object
for parsing URL query parameters into objects (#2647) authored by @GBrawl
In addition, http.send
has been enhanced to support caching overrides and in-band error handling (#2666 and #2187).
- Fix
opa build
to support custom built-in functions (#2738) authored by @gshively11 - Fix for file watching volume mounted configmaps (#2588) authored by @drewwells
- Fix discovery plugin to set last request and last successful request timestamps in status updates (#2630)
- Fix planner crash on virtual document iteration (#2601)
- Fix decision logger to requeue failed chunks (#2724 authored by @anderseknert)
- Fix object/set implementation in WASM-C library to avoid resizing.
- Fix JSON parser in WASM-C library to copy memory for strings and numbers.
- Improve WASM-C library to recycle object and set element structures while growing.
In addition, this release contains several fixes for panics identified by fuzzing:
- ast: Fix compiler to expand exprs in rule args (#2649)
- ast: Fix output var analysis to accept refs with non-var heads (#2678)
- ast: Fix panic during local var rewriting (#2720)
- ast: Fix panic in local var rewriting caused by object corruption (#2661)
- ast: Fix panic in parser post-processing of expressions (#2714)
- ast: Fix parser to ignore rules with args and key in head (#2662)
- ast: Fix object corruption during safety reordering
- types: Fix panic on reference to object with composite key (#2648)
- Renamed
timer_rego_builtin_http.send_ns
totimer_rego_builtin_http_send_ns
to avoid issues with periods in metric keys. - Removed deprecated
watch
package (#2265)
- Add support for H2C on HTTP listener (#2739 thanks @srenatus!).
- Add Go version information to
opa version
output (thanks @srenatus!) - The official OPA build has been updated to Go v1.14.9. Previously it was using v1.13.7 which is no longer supported (thanks @srenatus!)
This release contains a fix for a regression in v0.23.1 around bundle downloading. The bug caused OPA to cancel bundle downloads prematurely. Users affected by this issue would see the following error message in the OPA logs:
[ERROR] Bundle download failed: bundle read failed: archive read failed: context canceled
plugin = "bundle"
name = <bundle name>
- plugins/discovery: Set the last request and last successful request in discovery status (#2630)
- plugins/rest: Add response header timeout for REST client
The http.send
built-in function now supports caching across policy queries. The caching.inter_query_builtin_cache.max_size_bytes
configuration setting places a limit on the amount of memory that will be used for built-in function caching. By default, not limit is set. For http.send
, cache duration is controlled by HTTP response headers. For more details see the http.send
documentation.
OPA now supports a capabilities check on policies. The check allows callers to restrict the built-in functions that policies may depend on. If the policies passed to OPA require built-ins not listed in the capabilities structure, an error is returned. The capabilities check is currently supported by the check
and build
sub-commands and can be accessed programmatically on the ast.Compiler
structure. The repository also includes a set of capabilities files for previous versions of OPA under the capabilities/
directory.
For example, given the following policy:
package example
deny["missing semantic version"] {
not valid_semantic_version_tag
}
valid_semantic_version_tag {
semver.is_valid(input.version)
}
We can check whether it is compatible with different versions of OPA:
# OK!
$ opa build ./policies/example.rego --capabilities ./capabilities/v0.22.0.json
# ERROR!
$ opa build ./policies/example.rego --capabilities ./capabilities/v0.21.1.json
This release includes a new built-in function to test if a string is a valid regular expression: regex.is_valid
.
- Host environments no longer have to provide the
opa_println
function when instantiating compiled policy modules. - SDKs no longer have to set the heap top address during initialization.
- Add a new inter-query cache to cache responses across queries (#1753)
- Fix
opa
CLI flags to match documentation (#2586) authored by @OmegaVVeapon - Fix rule indexing when multiple glob.match mappers are required (#2617)
- Fix AST to marshal non-string object keys (#516)
- Fix signature calculation to include port if necessary (#2568)
- Fix partial evaluation to check function output for false values (#2573)
- Add
http.send
latency to query metrics (#2034) - Add support for
opa build
unknowns underdata
(#2581) - Add support to wait for plugin readiness before starting server
- Add parameter to set wall clock time during evaluation for replay purposes
- Fix groundness bit on objects during update
- Fix x509 built-in functions to parse PEM or DER inputs
- Fix bundle signing and verification to use standard JWT key ID header
- Optimize AST collections to cache hash values
- Optimize object iteration to avoid hashing
- Optimize evaluator by removing unnecessary term copying
-
The
watch
query parameter on the Data API has been deprecated. The query watch feature was unused and the lack of incremental evaluation would have introduced scalability issues for users. The feature will be removed in a future release. -
The
partial
query parameter on the Data API has been deprecated. Note, this only applies to thepartial
query parameter that the Data API supports, not Partial Evaluation itself. Thepartial
parameter allowed users to lazily trigger Partial Evaluation (for optimization purposes) during a policy query. While this is useful for kicking the tires in a development environment, putting optimization into the policy query path is not recommended. If users want to kick the tires with Partial Evaluation, we recommend running theopa build
command.
-
The
storage.Indexing
interface has been removed. Storage indexing has not been supported since 0.5.12. It was time to remove the interface. Custom store implementations that may have included no-op implementations of the interface can be updated. -
The
ast.Array
type has been redefined a struct. Previouslyast.Array
was a type alias for[]*ast.Term
. This change is backwards incompatible because slice operations can no longer be performed directly on values of typeast.Array
. To accomodate, theast.Array
type now exports functions for the same operations. This change decouples callers from the underlying array implementation which opens up room for future optimizations.
OPA now supports digital signatures for policy bundles. Specifically, a signed bundle is a normal OPA bundle that includes a file named ".signatures.json" that dictates which files should be included in the bundle, what their SHA hashes are, and of course is cryptographically secure. When OPA receives a new bundle, it checks that it has been properly signed using a key that OPA has been configured with out-of-band. Only if that verification succeeds does OPA activate the new bundle; otherwise, OPA continues using its existing bundle and reports an activation failure via the status API and error logging. For more information see https://openpolicyagent.org/docs/latest/management/#signing. Many thanks to @ashish246 who co-designed the feature and provided valuable input to the development process with his proof-of-concept #1757.
opa build
now supports multiple optimization levels. The first level (--optimize=1
) enables constant folding (based on partial evaluation) that only inlines values that can be computed entirely at build time. The second level (--optimize=2
) enables the existing (more aggressive) version of partial evaluation that eagerly inlines as much of the policy as possible. For more information on the optimization levels see the Optimization Levels section in the documentation.
numbers.range
(#2479) was added to support policies that need to generate a range of integers (e.g., a network port range).semver.is_valid
andsemver.compare
(#2538) was added to support policies that need to validate semantic version numbers (authored by @charlieegan3).
- All String built-in functions (except
sprintf
) are now implemented natively inside of Wasm-compiled policies.
- A few small issues in the Go integration and
rego
package examples have been resolved (#2294) and #2367) authored by @gaga5lala. - The Kubernetes Admission Controller tutorial as been updated to work with recent versions of Kubernetes (#2467 authored by @gaga5lala).
- A few issues in partial evaluation around negation inlining and partial rules have been resolved (e.g., #2492, #2491).
- OPA now supports IMDSv2 for the AWS metadata service. This improves the security posture of OPA deployments in AWS (#2482) authored by @nhw76.
- Several improvements to the project documentation including a policy style discussion, an integration option comparison, and discussion of bootstrapping and fail-open versus fail-closed modes.
- The project's CI/CD infrastructure has been migrated to GitHub Actions. The new CI/CD infrastructure is designed and implemented to be portable and includes a number of quality-of-life improvements.
- End-to-end query latency with decision logging enabled has been improved by 10%-15% in real-world cases.
- The
rego.Tracer
andrego.EvalTracer
API's have been deprecated in favor of the newerrego.QueryTracer
andrego.EvalQueryTracer
API. - The
tester.Runner#SetCoverageTracer
API has been deprecated in favor of the newertest.Runner#SetCoverageQueryTracer
API.
This release fixes #2497 where the comprehension indexing optimization produced incorrect results for nested comprehensions that close over variables in the outer scope. This issue only affects policies containing nested comprehensions that are recognized by the indexer (which is a relatively small percentage).
This release also backports the GitHub Actions migration and a fix to the Wasm library build step.
-
Decision log masks can now mutate decision log events. Previously, the masks could only erase data in the events. With this change, users can implement masks that obfuscate or add information to the decision log events before they are emitted. Thanks to @dkiser for implementing this feature #2379)!
-
This release contains a new built-in function for parsing X.509 Certificate Signing Requests (
crypto.x509.parse_certificate_request
). Thanks to @vivekbagade for implementing this feature #2402! -
This release adds support for aggregation and bit arithmetic operations for WebAssembly compiled policies. These functions no longer have to be provided by the host environment.
- cmd: Fix bug in --disable-inlining option parsing (#2196) authored by @Syn3rman
- docs: Improve terraform example to incorporate
child_modules
(#1772) - server: Fix panic caused by compiler misuse with bundles (#2197)
- topdown: Fix incorrect memoization during partial evaluation (#2455)
- topdown: Fix loss of precision in arithmetic and aggregate builtins (#2469)
-
Thanks to @Syn3rman for implementing an improvement to our release process to automatically tag external contributors (#2323)!
-
The coverage and profiling tracers no longer require variable values from the evaluator. This change improves perfomance significantly when coverage or profiling is enabled and policies inspect large data sets. Benchmarks show anywhere from 0.5x to over 30x speedup depending on the policy.
topdown.Tracer
has been deprecated in favor of a newer interfacetopdown.QueryTracer
.- All tracers (regardless of interface implementation) will now only be checked for being enabled at the beginning of query evaluation rather than on a per-event basis.
topdown.BuiltinContext#Tracers
has been deprecated in favor oftopdown.BuiltinContext#QueryTracers
. The olderTracers
field will benil
starting this release, and eventually removed.
- compile: Change name of result var for wasm binary (#2441)
- format: Deep copy inputs to avoid mutating the caller's copy (#2439)
- docs: Add
opa_println
to wasm required imports
- format: Refactor wildcard names to rewrite early (#2430)
- docs/content small output correction on terraform page (#1772)
- format: Fix wildcards in nested refs
- format: Fix panic with else blocks and comments (#2420)
This release fixes an issue in the Docker image build. The default ca-certificates were not being included becasue the Docker image is FROM scratch now.
This release includes a number of features, optimizations, and bugfixes.
OPA now determines the latest stable release version using
https://telemetry.openpolicyagent.org. The only information provided to the
telemetry service is the version (e.g., 0.20.0
), a UUIDv4 generated on
startup, and the build platform/architecture (e.g., darwin, amd64
). This
feature is on by default in opa run
however it can be easily disabled by
specifying --skip-version-check
on the command-line. If you are inside the
REPL, type help
to see the latest version information. If you are running OPA
as a server, OPA will log an INFO level message indicating if OPA is out of
date. Version checking is best-effort. Any errors that occur while communicating
with https://telemetry.openpolicyagent.org are only logged at DEBUG level. For
more information see https://openpolicyagent.org/docs/latest/privacy/.
The opa build
command can now be used to package OPA policy and data files
into bundles
that can be easily distributed via HTTP. See opa build --help
for details.
This change is backwards incompatible. If you were previously relying on opa build
to compile policies to wasm, you can still do so:
# before v0.20.0
opa build -d policy.rego 'data.example.allow'
# v0.20.0 and newer
opa build policy.rego -e example/allow -t wasm
This release includes a number of new built-in functions:
graph.reachable
for computing the transitive closure from edge sets. This function allows users to write policies that traverse organization charts, security groups, etc. (thanks to @jaspervdj-luminal!)io.jwt.verify_rs512
and other variants (rs
/es
/hs
/ps
,384
/512
) were added (thanks to @GBrawl!)uuid.rfc4122
for generating UUIDv4s (thanks to @reneklootwijk!)
This release also includes a few fixes to existing built-in functions:
units.parse_bytes
now supports units without theB
orb
suffix (thanks to @GBrawl!)io.jwt.verify_decode
now supports floating-pointnbf
andexp
claims (thanks to @GBrawl!)array.slice
clamping logic fixed to prevent panic (#2320).
-
The
opa run
command now supports a--diagnostic-addr
flag that causes the server to expose the/health
and/metric
endpoint on a different address. This makes it easier to secure sidecar deployments in Kubernetes because the main API endpoints can be served on localhost and the diagnostic endpoints can be served on 0.0.0.0 so that the kubelet and other components can access them (#2002). The envoy tutorial has been updated to show this in action. -
The AWS credential provided has been updated to support the standard
AWS_SESSION_TOKEN
andAWS_SECURITY_TOKEN
environment variables. These are used when signing S3 bundle requests for an AWS IAM assumed role (thanks to @kpiotrowski!)
This release includes a number of improvements for wasm compiled policies.
- UTF-8 and UTF-16 strings are now fully supported in the internal string representation (#1885)
- Numeric values are implemented on top of arbitrary-precision floating point numbers to avoid loss-of-precision issues.
- The arithemetic, set, array, and type checking built-in function categories are now supported by the wasm library. This means they do not have to be implemented by the language-specific opa-wasm SDKs.
- The set and object implementations now use a chained hash set under the hood (#2225)
- OPA will attempt to index collections generated by comprehensions to ensure linear runtime for policies performing "group-by" operations (e.g., inverting an objects.) For more information see the Policy Performance page (#2276).
- The OPA extension for VS Code now supports
Go To Definition
inside policies. This feature uses the newopa oracle find-definition
command. - The
opa test
command now includes location information on trace output. - The
opa fmt
command now preserveselse
block style when possible (thanks to @mikaelcabot!)
This release includes several improvements to the website and documentation.
- Improved terraform tutorial example (#1772) (thanks to @princespaghetti!)
- Fixed token validation logic in envoy tutorial example (#2395) (thanks to @princespaghetti!)
- Usability issues on the frontpage have been resolved (#2205, #2206 (thanks to @arunbsar!)
- The Policy Performance page now includes resource utilization guidelines (#1601)
- By popular demand, the "document model" explanation has been brought back into existence. It now lives in the Philosophy section (#2284).
- The Ecosystem page implements a simple sorting algorithm that ranks items by amount of related content.
- The policy cheat sheet has been merged into the Policy Reference page.
- REPL now correctly displays booleans in tabled output (#2338, thanks to @timakin!)
- Discovery now supports service configuration updates. This makes token refresh easier in distributed environments on AWS. (#2058)
- Fixed compiler panic if body omitted from
else
statement (#2353) - Fixed panic in /health API with the envoy plugin (#2396)
- Partial Evaluation no longer generates unsafe queries for certain negated expressions (#2045)
- Partial Evaluation no longer saves an incorrect binding list in some cases (#2368)
- Output variable analysis no longer visits closures. This makes the analysis easier to use outside of the safety check.
- Rules parsed from expressions now have location information set correctly.
- If you are building OPA for debian systems, the Makefile now supports a
make deb
target. The target requiresdpkg-deb
to be installed. Thanks to @keshto for contributing this! - OPA is now built, by default, with CGO disabled. Also, the default Docker
image (
openpolicyagent/opa
) is back to usingFROM scratch
.
-
An internal utility function that unmarshals JSON (
util.UnmarshalJSON
) has been fixed to return an error if the input bytes contain garbage following a valid JSON value. In the past, theutil.UnmarshalJSON
function would just return the valid JSON value and ignore the garbage following it. This change is backwards incompatible since clients that were previously transmitting bad data will now receive an error, however, we think it's important to surface errors rather than hide them (#2331). -
The Go plugin/shared library loading feature that was deprecated in v0.14.0 has finally been removed completely. If you are interested in extending OPA, see the Extensions for how to do so at compile-time (#2049).
-
The
github.com/open-policy-agent/opa/metrics#Counter
interface has been extended to require anAdd(uint64)
function. This change only affects users that have implemented their own version of thegithub.com/open-policy-agent/opa/metrics#Metrics
interface (which is the factory for counters.) -
As mentioned above, the
opa build
command-line syntax has changed. We think this is the right time to refresh the command and we are more confident that the new syntax will remain stable going forward.
-
This release deprecates
opa test -l
flag. Since we now display the trace with line information, this flag is no longer needed. -
In the next release we plan to deprecate the
?watch
and?partial
HTTP API parameters. The?watch
feature is unused and introduces significant complexity in the server implementation. The?partial
parameter lazily invokes Partial Evaluation inline with policy invocation. This is useful for development and debug purposes, however, it's not recommended for enforcement points ot use (since PE optimization can introduce significant latency.) Users should rely on the newopa build
command to perform PE on their policies. Seeopa build --help
for more information.
- plugins: Fix race between manager and plugin startup (#2343)
- cmd/fmt: Only list files if there were changes (#2295)
This release includes a new parser implementation that resolves a number of existing issues with the old parser. As part of implementing the new parser a small number of backwards incompatible changes have been made.
The new parser contains a small number of backwards incompatible changes that correct questionable behaviour from the old parser. These changes affect a very small number of actual policies and we feel confident in the decision to break backwards compatibility here.
- Numbers no longer lose-precision #501
- Leading commas do not cause objects to lose values #2198
- Rules wrapped with braces no longer parse #2199
- Rule names can no longer contain dots/hyphens #2200
- Object comprehensions now have priority over logical OR in all cases #2201
In addition there are a few small changes backwards incompatible changes in APIs:
- The
message
field onrego_parse_error
objects contains a human-readable description of the parse error. The old parser would often report "no match found" to indicate the input contained invalid syntax. The new parser has slightly more specific errors. If you integrated with OPA and implemented error handling based on the content of these human-readable error message strings, your integration may be affected. - The
github.com/open-policy-agent/opa/format#Bytes
function has been removed (it was unused.)
The output below shows the Go benchstat
result for master (5a5d2a42) compared to the new parser.
name old time/op new time/op delta
ParseModuleRulesBase/1-16 210µs ± 1% 4µs ± 1% -98.02% (p=0.008 n=5+5)
ParseModuleRulesBase/10-16 1.39ms ± 1% 0.03ms ± 0% -97.93% (p=0.008 n=5+5)
ParseModuleRulesBase/100-16 13.5ms ± 1% 0.3ms ± 1% -97.93% (p=0.008 n=5+5)
ParseModuleRulesBase/1000-16 148ms ± 5% 3ms ± 6% -97.77% (p=0.008 n=5+5)
ParseStatementBasicCall-16 141µs ± 5% 3µs ± 1% -97.92% (p=0.008 n=5+5)
ParseStatementMixedJSON-16 9.06ms ± 2% 0.07ms ± 1% -99.19% (p=0.008 n=5+5)
ParseStatementSimpleArray/1-16 131µs ± 6% 2µs ± 1% -98.10% (p=0.008 n=5+5)
ParseStatementSimpleArray/10-16 499µs ± 6% 7µs ± 2% -98.54% (p=0.008 n=5+5)
ParseStatementSimpleArray/100-16 4.00ms ± 2% 0.06ms ± 4% -98.58% (p=0.008 n=5+5)
ParseStatementSimpleArray/1000-16 42.0ms ± 3% 0.5ms ± 4% -98.70% (p=0.008 n=5+5)
ParseStatementNestedObjects/1x1-16 233µs ± 6% 4µs ± 3% -98.49% (p=0.008 n=5+5)
ParseStatementNestedObjects/5x1-16 514µs ± 0% 9µs ± 4% -98.33% (p=0.008 n=5+5)
ParseStatementNestedObjects/10x1-16 911µs ± 5% 14µs ± 5% -98.46% (p=0.008 n=5+5)
ParseStatementNestedObjects/1x5-16 4.24ms ± 1% 0.01ms ± 1% -99.82% (p=0.016 n=4+5)
ParseStatementNestedObjects/1x10-16 138ms ± 1% 0ms ± 1% -99.99% (p=0.008 n=5+5)
ParseStatementNestedObjects/5x5-16 714ms ± 0% 5ms ± 5% -99.26% (p=0.016 n=4+5)
ParseBasicABACModule-16 3.12ms ± 3% 0.04ms ± 4% -98.63% (p=0.008 n=5+5)
name old alloc/op new alloc/op delta
ParseModuleRulesBase/1-16 99.2kB ± 0% 5.7kB ± 0% -94.30% (p=0.008 n=5+5)
ParseModuleRulesBase/10-16 600kB ± 0% 29kB ± 0% -95.16% (p=0.008 n=5+5)
ParseModuleRulesBase/100-16 5.72MB ± 0% 0.27MB ± 0% -95.34% (p=0.008 n=5+5)
ParseModuleRulesBase/1000-16 58.0MB ± 0% 2.7MB ± 0% -95.42% (p=0.008 n=5+5)
ParseStatementBasicCall-16 70.2kB ± 0% 5.0kB ± 0% -92.82% (p=0.008 n=5+5)
ParseStatementMixedJSON-16 3.64MB ± 0% 0.06MB ± 0% -98.34% (p=0.008 n=5+5)
ParseStatementSimpleArray/1-16 63.7kB ± 0% 4.8kB ± 0% -92.42% (p=0.008 n=5+5)
ParseStatementSimpleArray/10-16 205kB ± 0% 8kB ± 0% -96.00% (p=0.008 n=5+5)
ParseStatementSimpleArray/100-16 1.64MB ± 0% 0.05MB ± 0% -97.19% (p=0.008 n=5+5)
ParseStatementSimpleArray/1000-16 16.5MB ± 0% 0.4MB ± 0% -97.50% (p=0.008 n=5+5)
ParseStatementNestedObjects/1x1-16 98.6kB ± 0% 5.7kB ± 0% -94.22% (p=0.008 n=5+5)
ParseStatementNestedObjects/5x1-16 224kB ± 0% 9kB ± 0% -96.05% (p=0.008 n=5+5)
ParseStatementNestedObjects/10x1-16 381kB ± 0% 13kB ± 0% -96.63% (p=0.008 n=5+5)
ParseStatementNestedObjects/1x5-16 1.76MB ± 0% 0.01MB ± 0% -99.38% (p=0.008 n=5+5)
ParseStatementNestedObjects/1x10-16 56.2MB ± 0% 0.0MB ± 0% -99.97% (p=0.008 n=5+5)
ParseStatementNestedObjects/5x5-16 280MB ± 0% 4MB ± 0% -98.67% (p=0.008 n=5+5)
ParseBasicABACModule-16 1.27MB ± 0% 0.04MB ± 0% -97.08% (p=0.008 n=5+5)
name old allocs/op new allocs/op delta
ParseModuleRulesBase/1-16 2.28k ± 0% 0.07k ± 0% -96.75% (p=0.008 n=5+5)
ParseModuleRulesBase/10-16 16.1k ± 0% 0.5k ± 0% -96.59% (p=0.008 n=5+5)
ParseModuleRulesBase/100-16 159k ± 0% 5k ± 0% -96.64% (p=0.008 n=5+5)
ParseModuleRulesBase/1000-16 1.62M ± 0% 0.05M ± 0% -96.72% (p=0.008 n=5+5)
ParseStatementBasicCall-16 1.36k ± 0% 0.05k ± 0% -96.25% (p=0.008 n=5+5)
ParseStatementMixedJSON-16 105k ± 0% 1k ± 0% ~ (p=0.079 n=4+5)
ParseStatementSimpleArray/1-16 1.34k ± 0% 0.04k ± 0% -97.09% (p=0.008 n=5+5)
ParseStatementSimpleArray/10-16 5.49k ± 0% 0.12k ± 0% -97.90% (p=0.008 n=5+5)
ParseStatementSimpleArray/100-16 47.8k ± 0% 0.8k ± 0% ~ (p=0.079 n=4+5)
ParseStatementSimpleArray/1000-16 481k ± 0% 8k ± 0% -98.33% (p=0.008 n=5+5)
ParseStatementNestedObjects/1x1-16 2.38k ± 0% 0.05k ± 0% -97.82% (p=0.008 n=5+5)
ParseStatementNestedObjects/5x1-16 6.02k ± 0% 0.12k ± 0% -97.94% (p=0.008 n=5+5)
ParseStatementNestedObjects/10x1-16 10.6k ± 0% 0.2k ± 0% -98.01% (p=0.008 n=5+5)
ParseStatementNestedObjects/1x5-16 51.2k ± 0% 0.1k ± 0% ~ (p=0.079 n=4+5)
ParseStatementNestedObjects/1x10-16 1.66M ± 0% 0.00M ± 0% -99.99% (p=0.008 n=5+5)
ParseStatementNestedObjects/5x5-16 8.16M ± 0% 0.07M ± 0% -99.13% (p=0.008 n=5+5)
ParseBasicABACModule-16 36.5k ± 0% 0.7k ± 0% -98.09% (p=0.008 n=5+5)
- ast: Add rules/functions that contain errors to the type env (#2155)
- ast: Fix panic when rule args contain call expressions (#2081)
- ast: Fix bug in term rewritten when 'input' is passed as an argument (#2084)
- bundle: Remove extra root name in bundle file ids (#2117)
- cmd/fmt: Fix to always write formatted file to stdout (#2235)
- cmd/test: --explain now turns on verbose output (#2069)
- cmd/test: Default
-v
traces show notes and fails (#2068) - docs/website: Fix mobile docs nav menu (#2074)
- format: Print var if wildcard is used multiple times (#2053)
- plugins/bundle: Update the downloader's e-tag based on bundle activation (#2220)
- plugins: Add support to specify bearer token path (which enables token refresh) (#2241)
- profiler: Fix panic when location is missing by grouping expressions missing a location (#2134)
- rego: Avoid re-using transactions in compiler (#2197)
- repl: Add unset-package command (#2140)
- server: Do not return partial modules /v1/policies output (#2036)
- server: Specify partial evaluation namespace to avoid conflicts (#2247)
- topdown: Add time.add_date builtin (#1990)
- topdown: Fix partial evaluation to save comprehensions correctly (#2243)
- topdown: Improve pretty trace location details (#2143)
- topdown: Include HTTP response headers in
http.send
output (#2238) - Multiple important improvements to
http.send()
thanks to @jpeach
- Added
man
target in the Makefile forman
page generation! (thanks to @olivierlemasle) - Added Sublime Text syntax file
- Added link to Emacs mode for Rego (thanks to @psibi)
- Added net.cidr_contains_matches built-in function
- Improved support for registering custom built-in functions
- Add
opa bench
andopa test --bench
sub commands for benchmarking policy evaluation. (#1424) - Permit verifying JWT's with a public key
http.send
improvements:- Allow for skipping TLS verification via
tls_insecure_skip_verify
option - Add
Host
header support
- Allow for skipping TLS verification via
- Bitwise operators (#1919)
bits.or
bits.and
bits.negate
bits.xor
bits.lsh
bits.rsh
json.remove
which works similar toobject.remove
but supports a JSON pointer path.
- docs: Render tutorials as list (#2071)
- ast: Fix type check for objects with non-json keys (#2183)
- ast: Return an error when parsing an empty module (#2054)
- docs: Fix broken PAM module link (#2113)
- docs: Fix code fence in kubernetes-primer.md (#2177)
- topdown: Invoke iterator when evaluating negation (#2142)
- Correct checkptr errors found with Go 1.14
opa parse
: fix panic when parsing invalid JSON
- The
ast.ParseModule
helper will now return an error if an empty module is provided. Previously it would return anil
error andnil
module. (#2054) - The
cmd
andtester
packages in OPA will now require Go 1.13+ to compile. Most library users should be unaffected.
- bundle: Dedicate
policy.wasm
for the compiled policy.
- vendor: Update xxhash to workaround checkptr errors with Go 1.14
- cmd/parse: Fix panic when parsing encounters an error
- Add location information into pretty printed trace output. (#2070)
- Add timeout for
http.send
builtin (#2099) - build: Force module mode and using only the vendor directory (#2063)
- cover: Exclude
some
expressions in coverage report (#1972) - docs: How to say "ray-go" (#2106)
- topdown: Make http.send() caching use full request (#1980)
- topdown: Wrap all builtin functions for errors normalization (#2101)
- topdown: http.send use provided CA without client certs (#1976)
- Add
object
manipulation built-ins - docs: Add link to Rego Playground in table of contents
- docs: Update tutorial with note about consistency
- topdown: Export builtin implementations outside the package
- ast: Fix rewriting vars in rule args (#2080)
-
This release improves partial evaluation to avoid saving statements when they do not depend on unknowns (e.g., comprehensions, references that require materializing the full extent of partial sets/objects, etc.) Also, expressions containing
with
statements are partially evaluated now. -
This release lets policy authors to de-reference calls (and other terms) without assigning the result to an intermediate variable. For example, instead of writing
a := f(x); a.foo == 1
users can now writef(x).a == 1
directly. Thanks @jaspervdj-luminal!
This release includes the following new built-in functions:
object.get
built-in function to lookup object keys with a fallback.crypto.md5
,crypto.sha1
,crypto.sha256
built-in functions to hash strings.
-
The
glob.match
built-in function was not defaulting the delimiter to "." like the documentation described. This was fixed in #2061 however the fix is not backwards compatible. If you are using theglob.match
built-in function, you should ensure that a delimiter is being supplied. A search of .rego files on GitHub only revealed a few instances of theglob.match
in-use so we decided to err towards fixing the broken behaviour rather than preserving buggy behaviour going forward. -
Related to the fix for #2031 and the changes with OPA v0.16.0 to use
/
separatedpath
's with the decision log plugin API. The decision logger will no longer modify theserver.Info#Path
field. Older versions would substitute.
for/
but this was causing incorrect results. As of v0.16.0 the server has been updated to provide the correct paths so REST API users are unaffected. Golang API users of theplugins.log.Logger#Log
interface may be impacted if passingast.Ref
style strings as a path as it will no longer be changed to/
separated. Callers need to do any transformation beforehand.
- docs: Update Kubernetes apiVersions to use
apps/v1
instead ofextensions/v1
(#1977) - plugins/logs: Leave the path unchanged for decisions (#2031)
- plugins/bundle: Include last successful request timestamp in status (#2009)
- plugins/bundle: Pass copy of status to bulk listeners (#1962)
- rego: Fix panic when partial evaluating with tracers (#2007)
- rego: Propagate custom builtins to
PartialResult
(#1792) - server: Update health check to use plugin status (#2010)
- topdown: Correct glob default delimeter (#2039)
- ast: Do not index expressions containing with statements
- ast: Fix panic in module parsing
- ast: Improve visitor performance by avoiding heap allocations
- cmd/build: return error if there is more than one positional argument
- docs: Fix live docs button size
- docs: Fix token validation example in Envoy tutorial
This release includes an important bugfix for users that enable tracing and use the "pretty" trace formatter.
- topdown: Fix bug in var rewriting during trace formatting (#2022)
- Fix for
*-rootless
Docker imagesUSER
being set incorrectly (#1982)
- Add
json.filter
to mask/filter nested fields (#1617) - Add
net.cidr_expand
to generate CIDR hosts
- Reduce server latency for indexed policies by ~30-40% by caching prepared queries across requests (#1958)
- Improve type checker error and trace output readability (#1430 and #1208)
- Re-create service clients to pickup certificate changes (#1898)
- Report full system path for bundle file locations (#1796)
- Add
status.console
option to log Status messages to console (#1937) - Fix
io.jwt.decode_verify
to support multiple keys in JWKS (#1901) - Fix
path
decision log field for queries against "/data" (1532)
This release also includes:
- Documentation improvements on how to use the
io.jwt.*
built-in functions for token verification - Metrics for bundle processing and activation (e.g., read, parse, and compile times)
- Better parse metric reporting in the server
- The fix for #1532 required a backwards incompatible change for v0 and v1
queries against "/data". This only affects queries against the exact path
"/data", not paths prefixed with "/data/", e.g., "/data/example/allow". Since
queries against "/data" are rare and normally only seen in development
environments, this is a low-impact change. As part of this change, the
server.Info#Path
field has been changed to use slash-separated paths instead of the string representation of Rego references (i.e., a dotted path rooted at "data"). If you are registering a logging callback directly against the server (e.g., by callingserver.Server#WithDecisionLoggerWithErr
) you will have to update your logging callback to deal with the new path format. Decision log consumers should treat a missing/emptypath
field as a query against "/data".
In this release we reached a milestone for Wasm: any Rego policy can be compiled to Wasm now! In the next few weeks we will focus on expanding on the set of built-ins supported out-of-the-box and inside the NodeJS SDK.
- bundle: Make the DirectoryLoader public (#1840)
- topdown: Add raw_body parameter to http.send (#1903)
- wasm: Update planner to support with keyword (#1116)
- ast: Fix NoWith helper on exprs
- ast: Fix module JSON unmarshalling
- build: Fix build-release.sh by removing obsolete make deps command
- docs: Add JWT verification examples to reference
- docs: Fix introduction to refer to rules consistently
- topdown: Add environment variable to dump tests to disk
- topdown: Provide rewritten query vars in traces
- types: Fix constant select on array types
- wasm: Add support for built-in functions
- wasm: Extend wasm library to support shallow copying
- wasm: Fix JSON string lexing and parsing
- wasm: Fix object insertion operation
- wasm: Fix opa_json_dump to terminate keywords properly
- wasm: Fix opa_set_add to set next element correctly
- wasm: Fix planner to check call expression for false return value
- wasm: Fix planning of virtual document extent
- wasm: Improve calling convention of eval() function
- wasm: Improve planner to reuse local variables
- wasm: Fix planner to plan default rule bodies
- wasm: Remove unnecessary condition statements for scans
- wasm: Store parsed numbers as strings
- website: Replace homepage with new version
This release includes many small improvements and bug fixes.
This release includes a few new built-in functions for string manipulation:
trim_left
,trim_right
,trim_prefix
, andtrim_suffix
(thanks @hasit)regex.find_all_string_submatch_n
andstrings.replace_n
(thanks @kenfdev)
- tester: Fix --timeout to apply to each test case (#1788)
- ast: Check for undefined functions before safety check (#1141)
- ast: Fix object corruption during local rewrite (#1852)
- ast: Fix virtual predicate used for rule index build (#1863)
- discovery: Fix log level message when on HTTP 304 (#1826)
- docs: Update Kubernetes primer test to avoid false-positives (#1794)
- repl: Fix unknown argument processing (#1670)
- topdown: Fix namespacing to use caller bindings (#1814)
- topdown: Fix units.parse_bytes implementation to use int64 (#1815)
- topdown: Fix base document dereference with composite (#1057)
- wasm: Add support for comprehensions (#1120)
- wasm: Add support for full virtual document model (#1117)
- wasm: Remove memory.grow calls on every malloc (#1121)
- build: Migrate to Go modules from Glide for dependency management
- build: Fix *-debug docker images to be ":debug" tag based
- ast: Replace "var" with "some" in SomeDecl#String
- ast: Add map of rewritten vars to Compiler
- topdown: Add API to disable rule indexing for evaluation
- bundle: Add more details to manifest root errors
- bundle: Ensure data paths use
/
separators for key - bundle: Fix for overwriting data file keys
- cmd: Ensure all errors are in JSON formatted CLI output
- cmd: Add source output format for partial eval
- cmd: Fix opa eval to specify profiler tracer correctly
- discovery: Support
resource
configuration option - rego: Don't propagate non-threadsafe fields from Rego to preparedQuery
- topdown: Fix namespacing to use caller bindings (#1814)
- file/loader: Standardize on forward slash paths
- Fix a number of links in the OPA documentation.
- Fix issue with bundle root path comparisons on Windows.
This release includes a large number of improvements to the docs as
well as performance optimizations that improve several end-to-end
benchmarks by ~25%. Also, the opa eval
and other sub-commands now
accept a -b
or --bundle
flag that tell OPA to treat file paths as
bundles (either .tar.gz or directories). This improves behaviour in
large or mixed workspaces.
- Status API messages now include a dump of OPA's Prometheus metric registry. This increases the Status API message size significantly (~6KB). If you are indexing the the Status API messages, consider removing the metrics. Nonetheless, for Status API implementations, having access to the Prometheus metrics is important for monitoring the health of the OPAs.
This release includes a few improvements to built-in functions:
-
A new function for converting SI strings (e.g., "10MB") to numbers:
units.num_bytes(x)
(#1561). This is useful in the context of Kubernetes if you need to deal with resource limits and requests. -
The
io.jwt.verify_*
functions have been extended to support JWKs.
This release also improves support for providing custom built-in functions to OPA. See the extensions documentation on openpolicyagent.org.
- ast, rego: Refactor unsafe built-in handling (#1666)
- ast: Fix ordering of rule type checking errors (#1620)
- ast: Update rule head to track assignments (#1541)
- ast: Fix bug that allowed recursion in dynamic refs (#1565)
- ast: Fix parsing of var-like scalars (#1582)
- docs: Add note about benchmark result page (#1275)
- docs: Update to show undefined example with != (#1626)
- docs: Update to use live blocks (#1650)
- format: Fix formatter to start line after writing comments (#1560)
- loader: Update to accept file:// URLs. (#1505)
- server: Improve decision log-related error messages (#1367)
- Add support for fuzzing the ast package in CI
- Add search bar powered by Algolia to the docs
- Add "type" field to decision log events sent to the console
- Add support for := assignments at file level
- Add build commit and version to runtime info
- Fix moduleLoader to copy returned parsed Modules
- Fix panic in /health?bundle=true
- Update the --plugin-dir flag as deprecated
- Update formatter to preserve rule assigmemnts
- Update metrics object to be thread-safe
- Support loading bundles and files w/ Rego API
- Fix panic in OPA HTTP server with
/health?bundle=true
when using bundles loaded from CLI (#1703).
- Fix panic in OPA HTTP server caused by concurrent map writes (#1666)
- Fix bundle plugin to report error in case bundle manifest roots overlap (#1635)
This release updates OPA to use the latest stable Golang release (1.12.8) that includes important fixes in the net/http package. See this golang-nuts group message for details.
This release adds support for downloading multiple bundles to OPA
using the new bundles
key in the configuration. APIs that include
bundle information have been updated to support multiple bundles:
- Status API messages include the status and revision of each bundle.
- Decision Log API messages include the revision of each bundle.
- Data API responses include the revision of each bundle in the provenance field if requested.
- Health API waits for all bundles to activate if requested.
These changes are backwards compatible. If you are using the
existing bundle
key in the configuration, you will not see any
changes in the APIs listed above.
We recommend that you switch to the new bundles
key and update
consumers of the above APIs to support multiple bundles.
For more information on bundles see the this page in the OPA documentation.
This release adds support for emitting decision logs to stdout. This is useful for shipping decision logs directly to existing logging backends.
You can enable console decision logging on the command line:
opa run --server --set decision_logs.console=true
Console decision logging can be enabled alongside normal and custom decision logging.
- ast: Report safety errors on line where expression starts (#1497)
- ast: Update rule index to support glob.match (#1496)
- bundle: Add support for loading YAML files from bundles (#1471)
- bundle: Cache compiler on storage context (#1515)
- cmd: Fix double print of rego errors (#1518)
- docs: Add section on how to express "FOR ALL" in Rego (#1307)
- docs: Fix mention of reference head var (#1477)
- docs: Remove cast_xyz functions from docs (#1405)
- server: Pass transaction in decision log event (#1543)
- storage: Add safety checks to in-memory store (#1594)
- topdown: Fix corrupt object panic caused by copy propagation (#1177)
- topdown: Fix virtual cache to allow composite key terms (#1197)
- OPA sets the User-Agent header in requests made to services.
openpolicyagent/opa:edge
Docker images are available now. Theedge
tag refers to the tip of master.- OPA supports signing and encoding of JWTs. See Token Signing for details.
- Prometheus metrics include cancelled HTTP requests.
- Compiler exposes optional unsafe built-in function check.
- Discovery query can be configured now. See Discovery Configuration for details.
- Optimized rewriteDynamics stage in compiler to reduce allocations.
- OPA subcommands support "fails" explanation now. The "fails" explanation is similar to the "notes" explanation except that it prints Fail events instead of Note events. This is useful for among other things, debugging test failures.
- Partial evaluation can disable inlining on specific virtual documents. If set correctly this can improve partial evaluation performance significantly because OPA can avoid computing cross-products.
rego.Rego#PrepareForEVal
now times partial evaluation properly.- The diagnostics feature deprecated in v0.10.1 has been removed.
- Fix performance impact of bundle activation on policy queries (#1516)
- Fix log masking to use correct transaction (#1551)
- Fix deadlock caused by log masking decision evaluation (#1543)
- Add decision log event for undefined decision on
POST /
endpoint
This release includes two new features and an important bug fix.
This release includes an important feature for protecting sensitive
information in decision logs: masking. With the new decision log
masking feature you can configure OPA to remove sensitive information
from the input
and result
fields of decision log events. See the
Decision Log documentation for details.
This release adds support for signing bundle download requests using an AWS signing scheme. This feature allows you to configure OPA to download bundles directly from S3. See the Configuration documentation for details.
- server: Fix deadlock caused by leaked write transaction (#1478)
- server: Add request headers to authorization input (#1456)
- rego: Add time zone support to time/date built-in functions
- eval: Add --instrument flag for profiling evaluation via command line
This release includes a few small but backward incompatible changes:
-
The compiler will reject functions that redeclare arguments. A search of public .rego files on GitHub only returned one result which was contained in the OPA documentation. For example:
f(x) { x := 1 # bad: redeclaration of 'x' x == 1 # ok }
-
Errors returned by built-in calls are no longer coded as
eval_internal_error
. Instead they are returned aseval_builtin_error
. This change is made so callers can differentiate between actual internal errors and built-in errors that are result of bad inputs from the policy. -
The
ast.QueryCompiler#WithInput
function andast.QueryContext#Input
field have been removed because they were unused and had no affect. -
The
ast.Compiler
andast.QueryCompiler
functions to register extra changes now require a stage and metric name.
This release includes a few notable features and improvements:
-
The
some
keyword allows you to declare local variables to avoid namespacing issues. See the Some Keyword section in the documentation for more detail. -
The
opa test
,eval
, REPL, and HTTP API have been extended with a new explanation mode for filtering tracing notes. This makes it easier to see the output oftrace(msg)
calls from your policy. -
The WebAssembly (Wasm) compiler has been extended to include support for compiling rules into Wasm. Previously the compiler relied on partial evaluation to inline all rules. In some cases this is not possible due to limitations on Rego queries. In coming releases, the Wasm support will be extended to cover the entire language.
-
The
rego
package has been extended to support prepared queries. Prepared queries cache the parsed and compiled query ASTs for re-use across multipleEval
calls. For small policies the speedup can be significant. See the GoDoc for details.
- Add Kubernetes admission control debugging tips (#1039)
- Add docs on health check API endpoint (#1086)
- Add hardened configuration example to security page (#1172)
- Add support for with keyword stacking (#802)
- Fix type inferencing on object keys (#1361)
- Fix simple Kubernetes deployment example (#874)
- Fix bug in data mocking that resulted in wrong iteration behavior (#1261)
- Fix bug in set deep copy that caused panic (#1406)
- Fix bug in REPL that prevented rules from being declared (#1104)
- docs: Better documentation for providing the
input
document over HTTP (#1293) - docs: Add note about HTTP_PROXY friends (#1410)
- Add CLI config overrides and ENV injection
- Add additional compiler metrics for each stage
- Add an “edge” release to the docs
- Add param to include bundle activation in /health response
- Add provenance query output
- Add support for graceful shutdown of OPA server
- Improve discovery feature documentation
- Make
json
logs the default and addjson-pretty
- Raise error when loading empty module in bundle
- Return eval_builtin_error instead of eval_internal_error
- Rewrite == to = in queries passed to the compile API
- docs: Update bundle docs with caching info
- Update logrus to 1.4.0
- server: Add early exit on PUT /v1/policies
- topdown: Fix set unification partial eval bug
- topdown: Omit rule body from enter/redo events
This release publishes the Hugo-based documentation to GitHub Pages 🎉
- Add
array.slice
built-in function (#1243) - Add
net.cidr_contains
andnet.cidr_intersects
built-ins (#1289). This change deprecates the oldnet.cidr_overlap
built-in function. The latter will be supported for backwards compatibility but new policies should refer tonet.cidr_contains
.
- Bump kube-mgmt container version to 0.8 in tutorial
- Remove unnecessary resizing allocs from AST set and object
- Add Kubernetes Admission Control guide
This release migrates the OPA documentation over to Hugo (from GitBook). Going forward the OPA documentation will be generated using Hugo and hosted on Netlify (instead of GitHub Pages). The Hugo/Netlify stack brings us inline with the goal for other CNCF projects and provides nice features like "preview before merge".
This release includes a small but backwards incompatible change to the
http.send
built-in. Previously, http.send
would always decode
responses as JSON even if the Content-Type was unset or explicitly not
JSON. If you were previously relying on HTTP responses that did not
set the Content-Type correctly, you will need to update your policy to
pass "force_json_decode": true
as in the http.send
parameters.
- Fix panic in mod operation (#1245)
- Fix eval tree enumeration to return errors (#1272)
- Fix http.send to handle non-JSON responses (#1258)
- Fix backticks in SSH example that were causing problems (#1260)
- Fix IAM examples to use regex instead of glob syntax (#1282)
- Add support to register custom stages in the compiler
- Add rootless Docker image stream
- Improve hash distribution on objects
- Reduce number of allocs in set membership implementation
- docs: Add homebrew install instruction to the Getting Started tutorial
- docs: Many improvements around := vs ==, best practices, cheatsheet, etc.
- cmd: Add --fail-defined flag to eval subcommand
- server: Fix patch path escaping
- These release contians a small but backwards incompatible change to the custom decision logger API. Custom decision loggers can now return an error which will cause the OPA to fail-closed.
- Fix substring built-in bounds checking (#1235)
- Add trailing newlines when pretty printing API responses
- Add default Go metrics to Prometheus
- Add pprof endpoint to HTTP server
-
This release adds support for scoping bundles to specific roots under
data
. This allows bundles to be used in conjunction with sidecars likekube-mgmt
that load local data and policy into OPA. See the Bundles page for more details. -
This release includes a small but backwards incompatible change to the Decision Log event format. Instead of including the OPA version as a top-level field, the OPA version is included in the labels. The OPA version field was only added in v0.10.3 so this should not impact many consumers.
- Add coverage support to
opa eval
sub-command - Fix path checking in server to prevent overlapping base and virtual docs (#1207)
- Fix cmd integration tests to cleanup plugin directory (#1185)
- Improve TLS support in
http.send
([#1067](https://github.com/open-policy-agent/opa/issues/
-
This release includes support for authentication via client certificates (thanks @srenatus!) For improvements to authentication see #1163.
-
This release includes a backwards incompatible change to the plugin interface. Specifically, when plugins are registered, callers must provide a factory that can validate configuration before instantiating the plugin. This allows OPA to ensure that all configuration is valid before activating changes. Since plugins were undocumented prior to this release, this change should be low impact. For details on plugin development see the new Plugins page on the website.
-
This release includes a backwards incompatible change to the HTTP decision logger event type. Specifically, "null" inputs are now handled correctly and decision logs for ad-hoc queries now populate the "query" field in the event instead of the "path" field. If you are using consuming decision log events in Go, please switch to the decision logger framework documented here: https://github.com/open-policy-agent/opa/blob/master/docs/book/plugins.md.
- Add OPA version to decision logs (#1089)
- Add query metrics to decision logs (#1033)
- Add health endpoint to HTTP server (#1086)
- Add line of failure in
opa test
(#961) - Fix panic caused by assignment rewriting (#1125)
- Fix parser to avoid duplicate comments in AST (#426)
- Fix semantic check for function references (#1132)
- Fix query API to return 4xx on bad request (#1081)
- Fix incorrect early exit from ref resolver (#1110)
- Fix rewriting of assignment values (#1154)
- Fix resolution inside references (#1155)
- Fix '^' location of lines starting with tabs (#1129)
- docs: Update count function doc to mention strings (#1126) (#1122)
- Add tutorial for OPA/Ceph integration using Rook
- Add metrics timer for server handler
- Add support for custom backends in decision logger
- Fix find operation on sets for non-empty refs
- Fix bug in local declaration rewriting
- Fix discovery docs to show a realistic example
- Update decision log event to include error
- Update decision log events to model paths and queries
- Update server and decision logger to represent input properly
- Update server to include decision ID in error events
- Avoid zero values in http.Transport{} in REST client
- wasm: Add support for composite terms (#1113)
- wasm: Add support for not keyword (#1112)
- wasm: Add == operator
- wasm: Add checks on single term and dot stmts
- wasm: Add support for boolean and null literals
- wasm: Add support for pattern matching on composites
- wasm: Fix planner for chained iteration
- wasm: Fix pretty printer writer usage
- wasm: Output filenames in testgen errors
- wasm: Refactor assignment for better typing
- wasm: Remove module dumping from build command
- wasm: Rename ir.LoopStmt to ir.ScanStmt
- wasm: Update tester to allow for missing cases
- Add manifest metadata to bundle data (#1079) (#1062)
- Add profile command to REPL (#838)
- Add decision ID note in API docs (#1061)
- Fix formatting of trailing comments in composites (#1060)
- Fix panic caused by input being set incorrectly (#1083)
- Fix partial eval to apply saved terms (#1074)
- Add Stringer implementation for expr values
- Add Stringer implementation on metrics object
- Add helper function to compile strings
- Add note to configuration reference about -c flag
- Add support for configuration discovery
- Add support for multiple tracers
- Add trace helper to rego package
- Add code coverage percentage
- Fix REPL to check number of assignment operands
- Fix bug in test runner rule name dedup
- Fix security link in REST API reference
- Fix formatting of empty sets
- Fix incorrect reporting of module parse time
- Fix out of range errors for eq/assign in compiler
- Fix parser to limit size of exponents
- Update compiler to iterate over modules in sort order
- Update OPA front page
- Mark diagnostics feature as deprecated
- Add show debug command to REPL (#750)
- Add
glob
built-ins for easier path matching (thanks @aeneasr) - Add support for specifying services as object
-
Wasm compiler. This release adds initial/experimental support for compiling Rego policies into Wasm executables. Wasm executables can be loaded and executed in compatible Wasm runtimes like V8 (nodejs). You can try this out by running
opa build
. -
Data mocking. This release adds support for replacing/mocking the
data
document using thewith
keyword. In the past,with
only supported theinput
document. This made it tricky to test context-dependent policies. With the newwith
keyword support, it's easier to write tests against contextual policies. -
Negation Optimization. This release includes an optimization in partial evaluation for dealing with negated statements (
not
keyword). In the past, OPA would generate a support rule for negated statements. This is harder for clients to consume and not readily optimized. The optimization computes the necessary cross-product of the negated query and inlines it into the caller. This leads to simpler partial evaluation results that are readily optimized, translated into other query languages (e.g., SQL and Elasticsearch), or compiled into Wasm.
- Add builtin to verify and decode JWT (#884)
- Add GoDoc sample for using rego.Tracer (#1002)
- Add built-in function to get runtime info (#420)
- Add support for YAML encoded input values (#290)
- Add support for client certificates (#684)
- Add support for non-zero exit code in eval subcommand (#981)
- Fix == rewriting on embedded terms (#995)
- Fix copy propagation panic in comprehensions (#1012)
- Implement regex.find_n (#1001) (#747)
- Improve with modifier target error (#343)
- Iterate over smaller set when intersecting (#531)
- Only write one trailing newline at end of file (#1032)
- Redirect HTTP requests with trailing slashes (#972)
- Update bundle reader to allow relative data.json (#1019)
- Expose version information via REST API (#277)
- Add default decision configuration
- Add extra helpers to loader result
- Add indentation to trace in failure output
- Add router option to the HTTP server
- Add support for headers in http.send (thanks @repenno)
- Deprecating --insecure-addr flag (thanks @repenno)
- Add POST v1/query API for large inputs (thanks @rite2nikhil)
- Remove heap allocations from AST set with open addressing
- Replace siphash with xxhash in AST
- Output traces on failures in verbose mode (thanks @srenatus)
- Rewrite duplicate test rule names (thanks @srenatus)
- Add option to enable http redirects (#921)
- Add copy propagation to support rules (#911)
- Add support for inlining negated expressions in partial evaluation
- Add deps subcommand to analyze base and virtual document dependencies
- Add partial evaluation support to eval subcommand
- Add
net.cidr_overlap
built-in function (thanks @aeneasr) - Add
regex.template_match
built-in function (thanks @aeneasr) - Add external security audit information (thanks @caniszczyk)
- Add initial support for plugin loading (thanks @vrnmthr)
- Fix copy propagator type assertion panic (#912)
- Fix panic in parser error detail construction (#948)
- Fix with value rewriting for call terms (#916)
- Fix coverage flag for test command (thanks @johscheuer)
- Fix compile operation timing in REPL
- Fix to indent 4 spaces instead of a tab (thanks @superbrothers)
- Fix REPL output in policy guide (thanks @ttripp)
- Multiple fixes in the Kubernetes admission controller tutorial (thanks @johscheuer)
- Improve formatting of empty ast.Body (#909)
- Improve Kubernetes admission control policy loading explanation (thanks @rite2nihkil)
- Update http.send test to work without internet access (#945)
- Update test runner to set Fail to true (#954)
- Improve token authentication docs and handler (#901)
- Link to security docs in tutorials (#917)
- Update bundle reader to cap buffer size (#920)
- Validate queries by checking unsafe builtins (#919)
- Fix XSS in debug page (#918)
- Add io.jwt.verify_es256 and io.jwt.verify_ps256 built-in functions (@optnfast)
- Add array.concat built-in function (#851)
- Add support for command line bundle loading (#870)
- Add regex split built-in function
- Fix incorrect AST node in Index events (#859)
- Fix terraform tutorial type check errors (#888)
- Fix CONTRIBUTING.md to include sign-off step (@optnfast)
- Improve save set performance (#860)
This release adds two major features to OPA itself.
-
Query Profiler: the
opa eval
subcommand now supports a--profiler
option to help policy authors understand the performance profile of their policies. Give it a shot and let us know if you find it helpful or if you find cases that could be improved! -
Compile API: OPA now exposes Partial Evaluation with first-class interfaces. In prior releases, Partial Evaluation was only used for optimizations purposes. As of v0.9, callers can use Partial Evaluation via HTTP or Golang to obtain conditional decisions that can be evaluated on the client-side.
- Add ADOPTERS.md file (#691)
- Add time.weekday builtin (#789)
- Fix REPL output for multiple bool exprs (#850)
- Remove support rule if default value is not needed (#820)
Here is a short list of notable miscellaneous improvements.
- Add any/all built-in functions (thanks @vrnmthr)
- Add built-in function to parse Rego modules
- Add copy propagation optimization to partial evaluation output
- Add docs for exercising policies with test framework
- Add extra output formats to eval subcommand
- Add support for providing input to eval via stdin
- Improve parser error readability
- Improve rule index to support unknown values
- Rewrite == with = in compiler
- Update build to enable CGO
...along with 30+ other fixes and improvements.
- Fix virtual document cache invalidation (#736)
- Fix partial cache invalidation for data changes (#589)
- Fix query to path conversion in decision logger (#783)
- Fix handling of pointers to structs (#722, thanks @srenatus)
- Improve sprintf number handling (#748)
- Reduce memory overhead of decision logs (#705)
- Set bundle status in case of HTTP 304 (#794)
- Add docs on best practices around identity
- Add built-in function to verify JWTs signed with HS246 (thanks @hbouvier)
- Add built-in function to URL encode objects (thanks @vrnmthr)
- Add query parameters to authorization policy input (#786)
- Add support for listening on a UNIX domain socket (#692, thanks @JAORMX)
- Add trace event for rule index lookups (#716)
- Add support for multiple listeners in server (thanks @JAORMX)
- Remove decision log buffer size limit by default
- Update codebase with various go-fmt/ineffassign/mispell fixes (thanks @srenatus)
- Update REPL command to set unknowns
- Update subcommands to support loader filter (#782)
- Update evaluator to cache storage reads
- Update object to keep track of groundness
- Add Kafka authorization tutorial
- Add URL query encoding built-ins
- Add runtime API to register plugins
- Update eval subcommand to support multiple files or directories (thanks @devenney)
- Update Terraform tutorial for OPA v0.8
- Fix bug in topdown query ID generation
This release includes a few major features that improve OPA's management capabilities.
-
Bundles: OPA can be configured to download bundles of policy and data from remote HTTP servers. This allows administrators to configure OPA to pull down all of the policy and data required at the enforcement point. When OPA boots it will download the bundle and active it. OPA will periodically check in with the server to download new revisions of the bundle.
-
Status: OPA can be configured to report its status to remote HTTP servers. The status includes a description of the active bundle. This allows administrators to monitor the status of OPA in a central place.
-
Decision Logs: OPA can be configured to report decision logs to remote HTTP servers. This allows administrators to audit and debug decisions in a central place.
The command line file loading convention has been changed slightly. If you were
previously loading files with opa run *
you should use opa run .
now. OPA
will not namespace data under top-level directory names anymore. The problem
with the old approach was that data layout was dependent on the root directory
name. For example opa run /some/path1
and opa run /some/path2
would yield
different results even if both paths contained identical data.
Thanks to @jyoverna for adding a trace
built-in function that allows policy
authors to include notes in the trace. For example, authors can now embed
trace
calls in their policies. When OPA encounters a trace
call it will
include a "note" in the trace. Callers can filter the trace results to show only
notes. This helps diagnose incorrect decisions in large policies. For example:
package example
allow {
input.method = allows_methods[_]
trace(sprintf("input method is %v", [input.method]))
}
allowed_methods = ["GET", "HEAD"]
- Add RS256 JWT signature verification built-in function (#421)
- Add X.509 certificate parsing built-in function (#635)
- Fix substring built-in bounds checking (#465)
- Generate support rules for negated expressions (#623)
- Ignore some built-in calls during partial eval (#622)
- Plug comprehensions in partial eval results (#656)
- Report safety errors for generated vars (#661)
- Update partial eval to check call args recursively (#621)
- Add base64 encoding built-in functions
- Add JSON format to test and check subcommands
- Add coverage package and update test subcommand to report coverage
- Add eval subcommand to run queries from the command line (deprecates opa run --eval)
- Add parse subcommand to parse Rego modules and print AST
- Add reminder/reminder (%) operator
- Update rule index to support ==
- Update to Go 1.10
- Various fixes to fmt subcommand and format package
- Fix input and data loading to roundtrip values. Allows loading of []string, []int, etc.
As well as many other smaller improvements, refactoring, and fixes.
- Use rego.ParsedInput to provide input from form (#571)
- Add omitempty tag for ad-hoc query result field
- Fix rego package to check capture vars
- Fix root document assignment in REPL
- Update query compiler to deep copy parsed query
-
Nested expressions: now you can write expressions like
(temp_f - 32)*5/9
! -
Assignment/comparison operators: now you can write
x := <expression>
to declare local variables andx == y
when you strictly want to compare two values (and not bind any variables like with=
). -
Prometheus support: now you can hook up Prometheus to OPA and collect performance metrics on the different APIs. (thanks @rlguarino)
This release adds and improves a bunch of new built-in functions. See the Language Reference for details.
- Add globs_match built-in function (thanks @yashtewari)
- Add HTTP request built-in function
- Add time.clock and time.date built-in functions
- Add n-way set union and intersection built-in functions
- Improve walk built-in function performance for partially ground paths
- Fix REPL assignment support (#615)
- Fix panic due to nil term value (#601)
- Fix safety check bug for call args (#625)
- Update Kubernetes Admission Control tutorial (#567)
- Update release script to build for Windows (#573)
- Add support for DELETE method in Data API (#609) (thanks @repenno)
- Add basic query performance instrumentation
- Add documentation covering how OPA compares to other systems
- Remove use of unsafe.Pointer for string hashing
This release adds initial support for partial evaluation. Partial evaluation allows callers to mark certain inputs as unknown and then evaluate queries to produce new queries which can be evaluated once inputs become known.
- Add initial implementation of partial evaluation
- Add sort built-in function (#465)
- Add built-in function to check value types
- Fix rule arg type inferencing (#542)
- Fix documentation on "else" keyword (#475)
- Fix REPL to deduplicate auto-complete paths (#432
- Improve getting started example (#532)
- Improve handling of forbidden methods in HTTP server (#445)
- Refactor sets and objects for constant time lookup
- Fix eval of objects/sets containing vars (#505)
- Fix REPL printing of generated vars
- Fix bug in expression formatting
- Fix dynamic rewriting to copy with modifiers
- Fix off-by-one bug in array helper
- Fix index usage for virtual docs (#490)
- Fix match error panic (#494)
- Fix wildcard mangling in rule head (#480)
- Add parse_duration_ns to generate nanos based on duration string
- Add product to calculate the product of array or set
- Fix docs example of set union
- Fix file watch bug causing panic in server mode
- Modify AST to represent function names as refs
- Refactor runtime to separate init and start
- Refactor test runner to accept Store argument
- Substitute comprehension terms requring eval (#453)
- Add alpine-based Docker image
- Add stdin mode to opa fmt
- Fix syntax error in comprehension example
- Improve input parsing performance in V0 API
- Refactor loader to read inputs once (allows use of process substitution)
- Remove backup creation from fmt subcommand
- Remove use of sprintf in formatter
This release adds a new test
subcommand to OPA. The test
subcommand enables policy unit testing. The unit tests are expressed as rules containing assertions over test data. The test
subcommand provides a test runner that automatically discovers and executes these test rules. See opa test --help
for examples.
- Fix type error marshalling bug (#391)
- Fix type inference bug (#381)
- Fix unification bug (#436)
- Fix type inferecen bug for partial objects with non-string keys (#440)
- Suppress match errors if closures contained errors (#438)
As part of this release, logrus was revendored to deal with the naming issue. If you use logrus, or one of your other dependencies does (such as Docker), be sure to check out sirupsen/logrus#570 (comment).
- Fix incorrect REPL interpretation of some exprs (#433)
- Fix inaccurate location information in some parser errors (#214)
- Add Terraform Testing tutorial to documentation
- Add shorthand for defining partial documents (e.g.,
p[1]
instead ofp[1] { true }
) - Add walk built-in function to recursively process nested documents
- Refactor Policy API response representations based on usage
This release adds Diagnostics support to the server. This greatly improves OPA's debuggability when deployed as a daemon.
- Fix data race in the parser extensions
- Fix index.html GET requests returning error on empty input
- Fix race condition in watch test
- Fix image version in HTTP API tutorial
- Add metrics command to the REPL
- Limit length of pretty printed values in the REPL
- Simplify input query parameters in data GET requests
- Update server to support pretty explanations
- Properly remove temporary files when running
opa fmt -d
- Add support for refs with composite operands (e.g,.
p[[x,y]]
)
- Add support for raw strings (#265)
- Add support to cancel compilation after some number of errors (#249)
- Add Kubernetes admission control tutorial
- Add tracing support to rego package
- Add watch package for watching changes to queries
- Add dependencies package to perform dependency analysis on ASTs
- Fix mobile view navigation bug
- Fix panic in compiler from concurrent map writes (#379
- Fix ambiguous syntax around body and set comprehensions (#377)
- Add support for set and object comprehensions
- Add support for system.main policy in server
- Add transaction support in rego package
- Improve type checking error messages
- Format REPL modules before printing them
- Correct
opa fmt
panic on missing files - Fix minor site issues
- Add rego examples with input and compiler
- Add support for query cancellation
OPA now supports user-defined functions that have the same semantics as built-in functions. This allows policy authors to quickly define reusable pieces of logic in Rego without overloading the input document or thinking about variable safety.
The storage layer has been improved to support single-writer/multiple-reader concurrency. The storage interfaces have been simplified in the process. Users can rely on https://godoc.org/github.com/open-policy-agent/opa/storage/inmem in place of the old storage package.
The website has been redesigned and the documentation has been ported over to GitBook.
OPA supports two new commands that check and format policies. Check out opa help
for more information.
- Add YAML serialization built-ins
- Add time built-ins
- Fixed incorrect source locations on refs and manually constructed terms. All term locations should be set correctly now.
- Fixed evaluation bug that caused partial sets and partial objects to be undefined in some cases.
This release includes a bunch of new built-in functions to help with string manipulation, JWTs, and more.
The JSON marshalling built-ins have been renamed. Policies that used json_unmarshal
and json_marshal
before will need to be updated to use json.marshal
and json.unmarshal
respectively.
- Add
else
keyword - Improved undefined built-in error message
- Fixed error message in
-
built-in - Fixed exit instructions in REPL tutorial
- Relax safety check for built-in outputs
This release includes a bunch of cool stuff!
- Basic type checking for queries and virtual docs (#312)
- Optimizations for HTTP API authorization policies (#319)
- New /v0 API to support webhook integrations (docs)
- Add support for namespaced built-ins (#314)
- Improve logging to include request/response bodies (#328)
- Add basic performance metrics (#320)
- Add built-ins to un/marshal JSON
- Add input form to diagnostic page
- Fix top-level navigation links
- Improve file loader error handling
- Fix recursive binding by short-circuiting ref eval (#298)
- Fix reordering for unsafe ref heads (#297)
- Fix rewriting of single term exprs (#299)
This release changes the run
command options:
- Removed glog in favour of Sirupsen/logrus. This means the command line arguments to control logging have changed. See
run --help
for details. - Removed
--policy-dir
option. For now, if policy persistence is required, users can treat policies as config files and manage them outside of OPA. Once OPA supports persistence of data (e.g., with file-based storage) then policy persistence will be added back in.
- Add support for additional HTTP listener (#289)
- Allow slash in policy id/path (#292)
- Improve request logging (#281)
- Add deployment documentation
- Remove erroneous flag.Parse() call
- Remove persist/--policy-dir option
- Replace glog with logrus
Also, updated Docker tagging so that latest points to most recent release (instead of most recent development build). The most recent development build can still be obtained with the {version}-dev tag.
This release adds support for TLS, token-based authentication, and authorization in the OPA APIs!
For details on how to secure the OPA API, go to http://openpolicyagent.org/documentation/references/security.
- Fix stray built-in error messages (#275)
- Update error codes and messages throughout (#237)
- Fix evaluation bug with nested value refs
- Fix rego.Eval to close transactions
- Fix buggy usage of errors.Cause
- Updated to support Go 1.8
- Fix issue in high-level Go API (#261)
- Fix parsing of inline comments (#258)
- Fix unset of input/data in REPL (#259)
- Handle non-string/var values in rule tree lookup (#257)
This release changes the Rego syntax. The two main changes are:
- Expressions are now separated by semicolons (instead of commas). When writing rules, the semicolons are optional.
- Rules are no longer written in the form:
head :- body
. Instead they are written ashead { body }
.
Also:
- Set, array, and object literals now support trailing commas.
- To declare a set literal with one element, you must include a trailing comma, e.g.,
{ foo, }
. - Arithmetic and set operations can now be performed with infix notation, e.g.,
x = 2 + 1
. - Sets can be referred to like objects and arrays now (#243). E.g.,
p[_].foo = 1 # check if element in has attr foo equal to 1
.
This release changes the evaluation behaviour for packages. Previously, if a package contained no rules OR all of the rules were undefined when evaluated, a query against the package path would return undefined. As of v0.4.2 the query will return an empty object.
This release changes the Data API to return an HTTP 200 response if the document is undefined. The message body will contain an object without the result
property.
- Allow sets to be treated like objects/arrays
- Added high level API for Go users. See
github.com/open-policy-agent/opa/rego
package. - Improved expression String() function to handle infix operators better.
- Added support for set intersection and union built-ins. See language docs.
For more details on these changes see sections in How Do I Write Policies.
- Added new default keyword. The default keyword can be used to provide a default value for rules with complete definitions.
- Added new with keyword. The with keyword can be used to programmatically set the value of the input document inside policies.
- Add basic REST API authorization benchmark
This release contains a few non-backwards compatible changes to the REST API:
-
The
request
document has been renamed toinput
. If you were calling the GET /data[/path]?request=value you should update to use POST requests (see below). -
The API responses have been updated to return results embedded inside a wrapper object:
{"result": value}
. This will allow OPA to return unambiguous metadata in future (e.g., pagination, analysis, etc.) If you were previously consuming Data API GET responses, you should update your code to access the value under the"result"
key of the response object. -
The API models have been updated to use snake_case (#222). This would only affect you if you were previously consuming error responses or policy ASTs.
The Data API has been updated to support the POST requests. This is the recommended way of supplying query inputs.
The built-in framework has been extended to support simplified built-in implementations:
- Refactor topdown built-in functions (#205)
- Add cURL note to REST API docs (#211)
- Fix empty request parameter parsing (#212)
- Fix handling of missing input document (#227)
- Improve floating point literal support (#215)
- Improve module parsing errors (#213)
- Fix ast.Number hash and equality
- Fix parsing of escaped strings
- Improve evaluation error messages
- Fixed unsafe vars with built-in operator names bug (#206)
- Fixed body to rule conversion bug (#202)
- Improved request parameter handling (#201)
- Improved release infrastructure
The last major/minor release of 2016! Woohoo! This release contains a few non-backwards compatible changes to the APIs.
These changes simplify and clean up the storage.Store interface. This should make it easier to implement custom stores in the future.
- Update storage to support context.Context (#155)
- Update underlying number representation (#154)
- Updates to use new storage.Path type (#159)
These changes update the language to align query arguments with state stored in OPA. With these changes, OPA can readily analyze policies and determine references that refer to state stored in OPA versus query arguments versus local variables.
These changes also update how query arguments are provided via the REST API.
- Updates to how query arguments are handled #197
- topdown.Context has been renamed to topdown.Topdown to avoid confusion with Golang's context.
- Add help topics to REPL (#172)
- Fix error handling bug in Query API (#183)
- Fix handling of prefixed paths with -w flag (#193)
- Improve exit handling in REPL (#175)
- Update parser support for = rules (#192)
- Add Visual Studio and Atom plugins
- Add lazy loading of modules during compilation
- Fix bug in serialization of empty objects/arrays
- Add YAML loading and refactor into separate file (#135)
- Add command line flag to eval, print, and exit (#152)
- Add compiler check for consistent rule types (#147)
- Add set_diff built-in (#133)
- Add simple 'show' command to print current module (#108)
- Added examples to 'help' output in REPL (#151)
- Check package declarations for conflicts (#137)
- Deep copy modules in compiler (#158)
- Fix evaluation of refs to set literals (#149)
- Fix indexing usage for refs with intersecting vars (#153)
- Fix output for references iterating sets (#148)
- Fix parser handling of keywords in variable names (#178)
- Improve file loading support (#163)
- Remove conflict error for same key/value pairs (#165)
- Support "data" query in REPL (#150)
- Add new compiler harness for ad-hoc queries
- Add tab completion of imports
- Added support for non-ground global values
- Added several new string manipulation built-ins
- Added TextMate grammar file
- Setup Docker image build and push to Docker Hub as part of CI
- Set literals
- Composite and reference values in Rule head
- Complete Rule definitions containing variables
- Builtins for regular expressions, string concatenation, casting to numbers
- Improved error reporting in parser and compiler
- Iteration over base and virtual documents (in the same reference)
- Query tracing and explanation support
- Pluggable data storage support
- GoDoc strings with examples
- REST API specification
- Concise language reference
- Per-query cache of virtual documents in topdown
And many other small improvements and fixes.
- Basic value types: null, boolean, number, string, object, and array
- Reference and variables types
- Array comprehensions
- Built-in functions for basic arithmetic and aggregation
- Incremental and complete rule definitions
- Negation and disjunction
- Module system
- Reference resolver to support packages
- Safety check to prevent recursive rules
- Safety checks to ensure successful evaluation will bind all variables in head and body of rules
- Initial top down query evaluation algorithm
- Basic in-memory storage that exposes JSON Patch style API
- REPL that can be run to experiment with ad-hoc queries
- Server mode supports HTTP APIs to manage policies, push and query documents, and execute ad-hoc queries.
- Basic build infrastructure to produce cross-platform builds, run style/lint/format checks, execute tests, static HTML site
- Introductions to policy, policy-enabling, and how OPA works
- Language reference that serves as guide for new users
- Tutorial that introduces users to the REPL
- Tutorial that introduces users to policy-enabling with a Docker Authorization plugin