Quick Start¶
Deploy with Helm¶
Add the Helm repository and install:
helm repo add kpod-metrics https://pjs7678.github.io/kpod-metrics
helm repo update
helm install kpod-metrics kpod-metrics/kpod-metrics \
--namespace kpod-metrics --create-namespace
Or install from a local clone:
git clone https://github.com/pjs7678/kpod-metrics.git
helm install kpod-metrics ./kpod-metrics/helm/kpod-metrics \
--namespace kpod-metrics --create-namespace
Try It Locally (kind)¶
Spin up a local demo cluster with a single command:
This creates a kind cluster, installs kpod-metrics, and sets up port-forwarding so you can immediately view metrics. Run ./scripts/quickstart.sh --cleanup to tear it down.
Verify¶
Check pods are running¶
You should see a pod per node with Running status.
Check health¶
kubectl -n kpod-metrics port-forward ds/kpod-metrics 9090:9090
curl http://localhost:9090/actuator/health | python3 -m json.tool
Check metrics¶
You should see metrics like kpod_cpu_context_switches_total, kpod_net_tcp_bytes_sent_total, etc.
View topology¶
Next Steps¶
- Configure collection profiles to control which metrics are collected
- Set up Grafana dashboards for visualization
- Enable Prometheus Operator integration
- Configure OTLP export for OpenTelemetry backends