Step 3: Configure and Deploy the Collector

Deploying the collector with Helm

The pganalyze collector runs as an ordinary container in your cluster. There’s nothing CNPG-specific about how the collector itself is deployed. It just needs network connectivity to the CNPG cluster’s -rw Service, the same way any other application in the cluster would connect.

Prerequisites

• kubectl is configured for your cluster
• The Helm CLI v3 or above is installed
• CNPG is already installed, and you have a Cluster with the monitoring role and pg_stat_statements set up from the previous two steps

Fetch the Helm chart’s example values

helm repo add pganalyze https://charts.pganalyze.com/
helm repo update
helm show values pganalyze/pganalyze-collector > collector-values.yaml

Below, you’ll edit collector-values.yaml to configure the collector, including your pganalyze API key and the connection details for the database you want to monitor.

We name it collector-values.yaml rather than the default values.yaml so it doesn’t collide with other charts’ values files in the same directory, such as Vector’s in the Log Insights setup.

Configure the connection settings

Update the extraEnv section of collector-values.yaml:

extraEnv:
  PGA_API_KEY: your-organization-api-key
  # PGA_API_BASE_URL: http://your-pganalyze-enterprise-server.internal
  DB_HOST: my-cluster-rw
  DB_NAME: app
  DB_USERNAME: pganalyze
  DB_PASSWORD: your_monitoring_role_password

1. The PGA_API_KEY can be found in the pganalyze Settings page for your organization, under the API Access tab
2. The PGA_API_BASE_URL is only needed if you’re using pganalyze Enterprise Server. Otherwise, leave it commented out.
3. The DB_HOST should be the CNPG -rw Service for your cluster: <cluster-name>-rw if the collector is deployed in the same namespace as the Cluster, or <cluster-name>-rw.<namespace>.svc if it’s in a different namespace. This Service always points at the current primary, so the collector keeps working across failovers without reconfiguration.
4. The DB_NAME is the database on the cluster you want to monitor. app is CNPG’s default application database name unless you customized .spec.bootstrap.initdb.database. To monitor more than one database on this cluster, use a comma-separated list instead, starting with the primary one, e.g. app,other_db, or end the list with * (e.g. app,*) to monitor every database on the cluster. See the collector settings reference for details
5. The DB_USERNAME and DB_PASSWORD are the credentials from the pganalyze-monitoring-user Secret you created in Step 1

extraEnv is fine for a first test, but since DB_PASSWORD is a credential, we recommend moving to a Kubernetes Secret-based config once you’ve confirmed things work. Save the following as pganalyze-collector-config.yaml:

apiVersion: v1
kind: Secret
metadata:
  name: pganalyze-secrets
stringData:
  pganalyze-collector.conf: |
    [pganalyze]
    api_key = your_api_key
    # api_base_url = http://your-pganalyze-enterprise-server.internal
    [my-cluster]
    db_host = my-cluster-rw
    db_name = app
    db_username = pganalyze
    db_password = your_monitoring_role_password

Then apply it:

kubectl apply -f pganalyze-collector-config.yaml

Finally, in collector-values.yaml, remove the extraEnv section and mount the Secret at /config instead:

volumes:
  - name: scratch
    emptyDir: {}
  - name: config-volume
    secret:
      secretName: pganalyze-secrets

volumeMounts:
  - mountPath: /tmp
    name: scratch
    subPath: tmp
  - mountPath: /state
    name: scratch
    subPath: state
  - mountPath: /config
    name: config-volume

Note: The pganalyze collector supports more optional settings (multiple database names, log forwarding via db_log_otel_server, etc.) that you can add to the [my-cluster] section above.

Install the Helm chart

helm upgrade --install my-collector pganalyze/pganalyze-collector --values=collector-values.yaml

Adjust the namespace if needed. You can run the same command again whenever you change collector-values.yaml.

To verify the install went well, check the deployment and pod logs:

$ kubectl get deployment
NAME                               READY   UP-TO-DATE   AVAILABLE   AGE
my-collector-pganalyze-collector   1/1     1            1           1m10s
$ kubectl logs deploy/my-collector-pganalyze-collector
I [my-cluster] Submitted compact snapshots successfully: 6 activity

The Submitted compact snapshots successfully message indicates the collector is running correctly.

Your setup is complete. The dashboard will start showing data within a few minutes.

Once you’ve confirmed the install is successful and you’re receiving query data in pganalyze, we recommend setting up Log Insights as a follow-up step, to automatically track log events in your database.

(Optional) Running tests

You can run the collector’s built-in test using docker-entrypoint.sh inside the running pod:

kubectl exec deploy/my-collector-pganalyze-collector -- /home/pganalyze/docker-entrypoint.sh test