Enable Prometheus monitoring¶

Type: How-To (Task-oriented)

Difficulty: Intermediate

Time: 15 minutes

This guide shows you how to expose Plone metrics to Prometheus by creating a ServiceMonitor for the backend, the frontend, or the Varnish cache.

Prerequisites¶

The Prometheus Operator installed in the cluster (it provides the ServiceMonitor CRD).
A Prometheus resource that selects the ServiceMonitor resources you create (check serviceMonitorSelector and serviceMonitorNamespaceSelector).
A Plone container instrumented to expose metrics over HTTP. Plone itself does not ship a Prometheus endpoint; you need an add-on (such as a WSGI middleware) or a sidecar exporter.

Enable the ServiceMonitor on the backend¶

import { Plone, PloneVariant } from '@bluedynamics/cdk8s-plone';

new Plone(chart, 'plone', {
  variant: PloneVariant.VOLTO,
  backend: {
    image: 'plone/plone-backend:6.1.3',
    servicemonitor: true,
    metricsPath: '/metrics',
  },
  frontend: {
    image: 'plone/plone-frontend:16.0.0',
  },
});

servicemonitor: true instructs cdk8s-plone to emit a ServiceMonitor that scrapes the backend Service on its main port at /metrics.

Scrape the frontend on a dedicated port¶

Volto can expose metrics on a separate port through middleware such as express-prometheus-middleware. Point cdk8s-plone at that port:

frontend: {
  image: 'plone/plone-frontend:16.0.0',
  servicemonitor: true,
  metricsPort: 9090,
  metricsPath: '/metrics',
}

metricsPort accepts a port number or a Service port name.

Scrape the Varnish cache¶

PloneHttpcache and PloneVinylCache each accept their own monitoring switch.

import { PloneHttpcache } from '@bluedynamics/cdk8s-plone';

new PloneHttpcache(chart, 'cache', {
  plone: ploneInstance,
  servicemonitor: true,
  exporterEnabled: true,
});

exporterEnabled (default true) deploys the Varnish exporter sidecar that the ServiceMonitor scrapes.

For the cloud-vinyl operator:

import { PloneVinylCache } from '@bluedynamics/cdk8s-plone';

new PloneVinylCache(chart, 'cache', {
  plone: ploneInstance,
  monitoring: true,
});

Verify the rollout¶

# Generate manifests and confirm the ServiceMonitor exists
cdk8s synth
grep -l 'kind: ServiceMonitor' dist/*.yaml

# Apply and inspect on the cluster
kubectl apply -f dist/
kubectl get servicemonitor -n <namespace>
kubectl describe servicemonitor <name> -n <namespace>

If Prometheus is not picking up the new target, confirm:

The ServiceMonitor namespace matches the Prometheus resource’s serviceMonitorNamespaceSelector.
The ServiceMonitor labels match the Prometheus resource’s serviceMonitorSelector.

The metrics endpoint returns HTTP 200 from a pod in the cluster:

kubectl run -it --rm curl --image=curlimages/curl --restart=Never -- \
  curl -sf http://<service>.<namespace>:<port>/metrics | head