Kubernetes Guide

BetterDB Monitor ships an official Helm chart. The chart deploys the monitor itself; the Valkey/Redis instance it watches (and the optional PostgreSQL for durable history) run wherever they already run — in-cluster, managed, or on a VM.

Run the betterdb-monitor Helm chart 

helm repo add betterdb https://docs.betterdb.com/charts
helm repo update
helm install betterdb-monitor betterdb/betterdb-monitor \
  --namespace betterdb --create-namespace \
  --set db.host=my-valkey.default.svc.cluster.local \
  --set db.password=yourpassword

Open the dashboard:

kubectl port-forward -n betterdb svc/betterdb-monitor 3001:3001
# http://localhost:3001

db.host can be omitted — the monitor starts without a connection and you add one in the UI. Inside a cluster, use the Kubernetes service DNS name of your database (<service>.<namespace>.svc.cluster.local), not localhost.

Override default values

Create a values.yaml with the values you want to override (see the annotated defaults in charts/betterdb-monitor/values.yaml), then pass -f values.yaml to helm install / helm upgrade. A typical production profile:

db:
  host: my-valkey.default.svc.cluster.local
  existingSecret: my-valkey-auth        # key: db-password

storage:
  type: postgres                        # durable monitoring history
  existingSecret: my-monitor-storage    # key: storage-url

ingress:
  enabled: true
  className: nginx
  hosts:
    - host: monitor.example.com
      paths:
        - path: /
          pathType: Prefix
  tls:
    - hosts: [monitor.example.com]
      secretName: monitor-tls

Persistent storage

Two kinds of state, two knobs:

  • Monitoring history (slowlogs, metrics, anomaly events) lives in the storage backend, not on the pod’s disk. The default storage.type=memory is ephemeral — history disappears on pod restart. Set storage.type=postgres and storage.url (or storage.existingSecret) for history that survives restarts and rescheduling.
  • License state (offline tokens, the online outage-grace token) is kept under /app/data. Enable persistence.enabled=true to back it with a PersistentVolumeClaim so paid tiers survive rescheduling. The chart sets fsGroup: 1001 to match the image’s non-root user, so no manual chown is needed (unlike Docker volumes).

Overriding configuration

Every environment variable from Configuration can be set through the chart. First-class values exist for the common ones (db.*, storage.*, license.*, telemetry); everything else goes under extraEnv:

extraEnv:
  - name: ANOMALY_PROMETHEUS_INTERVAL_MS
    value: "15000"
  - name: OTEL_EXPORTER_OTLP_ENDPOINT
    value: "http://otel-collector.observability:4318"

Secrets never need to be inlined: db.existingSecret, storage.existingSecret, license.existingSecret, and license.offline.existingSecret all reference Secrets you manage yourself (External Secrets, SealedSecrets, plain kubectl create secret, …).

Air-gapped licensing 

kubectl create secret generic betterdb-license -n betterdb \
  --from-file=license.jwt=/path/to/betterdb-license.jwt

license:
  offline:
    existingSecret: betterdb-license
persistence:
  enabled: true

With an offline token and no online key the monitor makes zero outbound requests. Details: Offline licenses.

Upgrade the helm chart 

helm repo update
helm upgrade betterdb-monitor betterdb/betterdb-monitor \
  --namespace betterdb --reuse-values

The default image tag is pinned to the chart’s appVersion (the -no-ai variant), so helm upgrade after a repo update moves you to the release the chart was published for — no :latest surprises. Release notes: Updating.