Skip to main content

​
Admin UI logs

  1. Open Deployments.
  2. Open the target deployment.
  3. Go to Logs.
  4. Select a pod.
  5. Use Reconnect to restart the stream or Export to download the captured output as <pod-name>-logs.txt.
The Logs tab streams pod logs in real time. If the stream disconnects, existing captured logs remain visible and the status badge changes to disconnected.

​
Admin UI metrics

  1. Open the target deployment.
  2. Go to Metrics.
  3. Select a non-Sentinel pod.
  4. Use Reconnect to restart the live metrics stream.
The Metrics tab builds live charts from pod metric snapshots. Current panels include CPU, memory, clients, network I/O, command rates, command latency, graph command latency, graph command rates, DB keys, and uptime. The Open in Grafana button appears only when settings.ui.grafana_url is configured.
  1. Open System Settings.
  2. Open the UI/settings section.
  3. Set the Grafana URL.
  4. Save settings.
  5. Return to deployment Metrics and select Open in Grafana.
When opening Grafana, the UI appends:
  • var-namespace=<namespace>
  • var-pod=<selected-pod>
  • var-pod_scope=<selected-pod>

​
Helm observability resources

The chart has these optional observability toggles: When podMonitor.enabled=true, the chart creates:
  • An Admin Server PodMonitor that scrapes /metrics on the http port.
  • A FalkorDB PodMonitor when the FalkorDB addon is enabled. It selects KubeBlocks-managed non-Sentinel pods and scrapes /metrics on http-metrics and graph-metrics.
PodMonitor and PrometheusRule require Prometheus Operator CRDs. GrafanaDashboard requires the Grafana Operator CRD.

​
Protecting the Admin Server metrics endpoint

/metrics is served on the same port as the API but outside the authenticated /api prefix, so by default anything that can reach the pod can read it. Set HTTP Basic credentials to close that:
The chart then creates a Secret named <release>-falkordb-enterprise-admin-server-metrics, wires it into the Admin Server, and adds a matching basicAuth block to the Admin Server PodMonitor, so podMonitor.enabled=true keeps scraping without further changes. Any scraper you configure yourself must send the same credentials, or it will get 401. To keep the credentials out of Helm values, create the Secret yourself with username and password keys and reference it:
The Admin Server refuses to start if only one of the username and password is set, so a partial configuration cannot silently leave the endpoint open.

​
Example values

Use a real Grafana dashboard JSON for production. The chart requires each configured dashboard to include name and json.

​
Validate

In the Admin UI, verify Logs, Metrics, and Open in Grafana from a deployment details page.