Monitoring

Observe service health with metrics, logs, and traces that share a resource context.

Three connected signals#

Metrics show trends, logs explain events, and traces reveal where request time is spent. Digital Universe attaches workspace, project, environment, and service labels to each signal so you can move between them without rebuilding a query.

Signal	Recommended starting point
Metrics	Request rate, error ratio, and duration
Logs	Structured messages with severity and request ID
Traces	Gateway-to-service request spans

Alert on sustained errors#

apiVersion: universe/v1
kind: AlertPolicy
metadata:
  name: catalog-error-rate
spec:
  service: catalog
  signal: http.error_ratio
  threshold: 0.02
  window: 5m
  minimumRequests: 100
  severity: warning

Make alerts actionable#

Every alert should identify a service owner and a runbook. A minimum request count avoids noisy ratios on low-traffic services. Test notification routing in development before enabling a production policy.