Skip to main content

Metrics

Phase Two provides built-in metrics for dedicated clusters through the self-service dashboard. Metrics give you aggregated, charted views of authentication activity and HTTP traffic so you can spot trends, errors, and performance issues at a glance.

Visit your cluster in the self-service dashboard and open the Metrics section. It is split into two views:

  • Event Metrics — authentication and admin activity derived from Keycloak events
  • Request Metrics — HTTP traffic, errors, caching, and latency

All charts share the same filter bar (time range, realm, and refresh controls) and all timestamps are in UTC.

Filters and controls

  • Time range presets: Last 15 minutes, Last 1 hour, Last 6 hours, Last 24 hours, Last 2 days, Last 7 days, and Last 30 days
  • Custom time window: set exact From and To values
  • Realm: limit results to a specific deployment/realm in the cluster, or view all realms
  • Refresh: rerun the current query immediately
  • Charts support zooming into a time window; use the reset control to restore the previous range

Event Metrics

Event Metrics chart Keycloak user and admin activity over the selected time range:

  • User events by type
  • Admin events by resource / operation
  • Successful logins
  • Failed logins
  • Active users per day
  • New registrations
  • Password reset requests
  • MFA enroll / remove
  • Login error reasons (lockouts)
  • Session concurrency (proxy)
Resources Overview

Request Metrics

Request Metrics chart HTTP traffic and performance for the cluster:

  • Requests by endpoint type
  • Requests by HTTP status class
  • Error rate (4xx + 5xx share)
  • Cache hit ratio
  • Latency p50 / p95 / p99 (time_taken, seconds)
  • Latency p95 by endpoint type
  • TTFB p50 / p95 / p99 (seconds)
  • Bytes served
  • Unique client IPs (approximate)
  • Top endpoints
  • Top user agents
Resources Overview

Access notes

If you see a message saying metrics cannot be viewed due to lack of permissions, your account does not currently have access to that cluster's metrics. See Logs for log-level investigation of individual entries.