Nango requires specific components and configurations to operate correctly within a Kubernetes cluster. Key dependencies include:
- Elasticsearch: Nango relies on Elasticsearch for logging.
Nango can operate without running Elasticsearch and the environment variable
NANGO_LOGS_ENABLED
determines if logs are ingested and displayed in the Nango UI. If you are running Elasticsearch you need to set:
NANGO_LOGS_ENABLED
NANGO_LOGS_ES_PWD
NANGO_LOGS_ES_URL
NANGO_LOGS_ES_USER
Note that these helm charts include a self hosted Elasticsearch version but using that for production workloads is not recommended.
- Required Values: Obtain the following values from a Nango developer:
MAILGUN_API_KEY
Nango expects the following secrets:
This secret should contain:
postgres-password
: Required ifpostgresql.enabled
is set tofalse
(i.e., using an external database).encryption-key
: Required ifshared.encryptionEnabled
is set totrue
.mailgun-api-key
.
Example command to create nango-secrets
:
kubectl create secret generic nango-secrets \
--from-literal=postgres-password=secure-pw \
--from-literal=encryption-key=base64-encoded-256-bit-key \
--from-literal=mailgun-api-key=key-from-nango-dev
Then to create the secrets:
apiVersion: v1
kind: Secret
metadata:
name: nango-secrets
type: Opaque
data:
postgres-password: [base64-encoded-password]
encryption-key: [base64-encoded-encryption-key-which-means-you-need-to-base64-encode-the-base64-256-bit-key]
mailgun-api-key: [base64-encoded-api-key]
To facilitate syncs and actions, Nango utilizes a persistent volume across services. The jobs-pvc.yaml and jobs-${aws|gcp}-storage-class.yaml in the jobs template are responsible for creating this volume, used by the jobs and server components. Reach out for assistance if you encounter issues with the volume attachment.
The server component, crucial for OAuth handshake, needs to be publicly accessible. By default, on AWS, a LoadBalancer exposes the server. Set useLoadBalancer to false to use an alternate exposure method.
Note for Porter Users: There's a known issue where the server might not be correctly exposed due to an internal-only load balancer. Please contact us for support.
-
Install helm: Follow the official Helm documentation
-
Add the Nango Repository
helm repo add nangohq https://nangohq.github.io/nango-helm-charts
- Update the Repository (if previously added):
helm repo update nangohq
helm search repo nangohq
- Configure values.yaml: Refer to the configuration section below.
- Install Nango charts
helm install nango nangohq/nango
- To uninstall the chart
helm delete nango
Component | Key | Default Value |
---|---|---|
postgresql | enabled | true |
fullnameOverride | nango-postgresql | |
primary.persistence.enabled | false | |
primary.resources.limits.cpu | "1000m" | |
primary.resources.limits.memory | "2048Mi" | |
primary.resources.requests.cpu | "250m" | |
primary.resources.requests.memory | "1024Mi" | |
auth.postgresPassword | nango | |
auth.database | nango | |
elasticsearch | enabled | false |
fullnameOverride | nango-elasticsearch | |
clusterName | elastic | |
security.enabled | true | |
security.elasticPassword | nango | |
server | name | server |
tag | enterprise | |
useLoadBalancer | true | |
replicas | 1 | |
jobs | name | jobs |
tag | enterprise | |
replicas | 1 | |
volume.name | flows-volume | |
volume.claimName | flow-claim | |
volume.aws | false | |
volume.gcp | false | |
runner | name | runner |
tag | enterprise | |
replicas | 1 | |
persist | name | persist |
tag | ||
replicas | 1 | |
url | http://nango-persist |
|
shared | namespace | default |
ENV | production | |
DB_HOST | nango-postgresql | |
DB_USER | postgres | |
DB_PORT | "5432" | |
DB_NAME | nango | |
DB_SSL | false | |
APP_URL | https://your-hosted-instance.com |
|
CALLBACK_URL | https://your-hosted-instance.com/oauth/callback |
|
flows_path | /flows | |
useVolumeForFlows | true |