Dashboards & Viewing Logs

Concepts are in the System Monitoring overview. This is a how-to.

The OpenG2P — Logs & Health dashboard

A ready-made dashboard is provisioned automatically. Open Grafana → Dashboards and search for "OpenG2P" (OpenG2P — Logs & Health).

It combines logs (Loki) and health (Prometheus) in one view:

Panel	Shows
Error lines (1h), CrashLooping pods, Total log rate, OOMKilled (1h)	At-a-glance health stats
Log volume by namespace	Where logs are coming from / spikes
Error rate by service	Which service is erroring
Top erroring pods (table)	The noisiest failures right now
Pod restarts (15m)	Crash/restart activity
Recent errors (logs)	Live list of error lines, click-through to detail

Screenshot: add an image at .gitbook/assets/system-monitoring-dashboard.png and embed it here.

Adding your own dashboards: drop a dashboard *.json into the automation's charts/grafana-dashboards/dashboards/ directory; it is imported automatically on the next deploy (no Grafana restart).

Viewing logs ad-hoc (Explore)

Use Explore (compass icon) → select the Loki datasource → write a query in LogQL. Set a time range (top-right) and click Run query.

{k8s_namespace_name="kube-system"}                 # all logs in a namespace
{service_name="esignet"}                           # one service
{k8s_namespace_name=~".+"} |~ `(?i)error|fatal`    # only errors, cluster-wide
{service_name="esignet"} | json | level="ERROR"    # parse JSON logs and filter

Useful labels: k8s_namespace_name, k8s_pod_name, k8s_container_name, k8s_deployment_name, service_name.

• Patterns tab — Loki auto-groups similar log lines into templates, so you can spot "this error happened 4,000×" without writing a query.

• Near-real-time: use the auto-refresh dropdown (e.g. 10s) for live-style updates.

Live tail (the streaming "Live" button) does not work through the Rancher proxy — it needs WebSockets, which the proxy blocks. Use a range query with auto-refresh instead.