zworkforce-observability

name	zworkforce-observability
description	Configure and validate zWorkforce OTLP/HTTP JSON tracing, Prometheus metrics, and the Grafana dashboard so operations reflect real task, agent, and provider activity.

zWorkforce Observability

Make platform behavior debuggable and SLOs enforceable.

Workflow

1. Identify the domain under review (task runtime, scheduler/outbox leadership, provider pool, workflow execution, or API layer) and its expected trace spans and metric names.
2. Verify OTLP/HTTP JSON trace export is enabled and spans cover the full task lifecycle: submission, lease, tool execution, and completion.
3. Confirm Prometheus metrics exist for the SLOs being enforced (latency, error rate, queue age, provider circuit state) and that the Grafana dashboard panels are wired to those exact metric names.
4. Flag any area with dashboard panels but no underlying metric, or metrics with no alerting/SLO tied to them, before declaring it observable.
5. Reproduce a real failure or slow path and confirm it is visible in traces and metrics before closing out the review.

References

• zworkforce/telemetry.py
• zworkforce/metrics.py
• docs/OBSERVABILITY.md
• deploy/observability

Output

Report instrumented spans/metrics, dashboard coverage, SLO alignment, and any blind spots that need additional instrumentation.