Skip to main content

Observability and Alerting

Observability determines whether you catch latency and reliability regressions early.

Signals to Track​

Minimum baseline:

  • session setup success/failure
  • handshake error rate
  • direct vs relay ratio
  • p95/p99 latency indicators
  • session drop/disconnect rates

Linux host focus:

  • portal preflight failures
  • PipeWire/session backend failures
  • host encoder/capture startup failures

Log Strategy​

Use structured logs and include:

  • timestamp with timezone
  • component (gateway, relay, host, client)
  • session/correlation identifiers when available
  • clear error category and action context

Do not log:

  • secrets/tokens/private key material
  • sensitive payload data

Alert Baseline​

Create alerts for:

  1. gateway health endpoint failure
  2. auth failure surge
  3. handshake failure surge
  4. relay registration/heartbeat anomalies
  5. sudden direct-to-relay ratio shift

Tune thresholds per environment tier.

Dashboard Baseline​

At minimum, maintain dashboards for:

  • control-plane health and request volume
  • session quality and adaptation trends
  • relay utilization and error states
  • Linux host runtime health markers

Release Observability Checklist​

Before rollout:

  1. baseline current metrics
  2. define rollback thresholds
  3. enable focused alerting during rollout window

After rollout:

  1. compare latency and failure deltas
  2. verify no hidden increase in relay usage
  3. confirm error distribution stability

Troubleshooting Correlation Tips​

When an issue appears:

  1. align client, host, and control-plane timestamps
  2. compare before/after deployment windows
  3. isolate whether change is network, runtime, or control-plane induced