Skip to content

Install with Helm

Install alertkube, point it at Slack, and verify health.

Prerequisites

  • Kubernetes access with kubectl.
  • Helm 3.8+.
  • A Slack incoming webhook URL.

Confirm your tools are ready:

kubectl cluster-info
helm version

Install

helm upgrade --install alertkube oci://ghcr.io/aryasoni98/charts/alertkube --version 1.2.0 \
  --set cluster=my-cluster \
  --set slack.webhookUrl=https://hooks.slack.com/services/Change-Me \
  --set slack.channels.critical=alerts-critical \
  --set slack.channels.warning=alerts-warning \
  --set slack.channels.info=alerts-info

cluster is shown in every alert. Use a name responders recognize.

Verify the Pod

Watch the deployment roll out:

kubectl get pods -l app.kubernetes.io/name=alertkube -w

If it does not become ready, inspect logs:

kubectl logs -l app.kubernetes.io/name=alertkube --tail=50

Check Health and Metrics

kubectl port-forward deploy/alertkube 9090:9090

In another terminal:

curl -s http://localhost:9090/healthz
curl -s http://localhost:9090/metrics | grep alertkube_

/healthz should return success, and /metrics should expose alertkube_* series.

Common first-run notes:

  • Modern Slack webhooks ignore per-channel routing. Use bot-token mode for real severity channels.
  • Set cluster; otherwise alerts use the placeholder.
  • Node alerts require cluster-scoped RBAC.
  • OCI installs should include --version.

Next steps