Metrics, traces and logs

Available in 0.1: /metrics (Prometheus), OTLP trace export with one span per run, JSON logs with the run ID, /healthz and /readyz. Planned for 0.2: spans per step, model and tool call and metrics for tool calls, approvals, tokens and budget exhaustion.

• Metrics: /metrics in Prometheus format: runs by status, step durations, tokens and cost by agent and provider, gate decisions by reason, control agent actions, queue depth.
• Traces: OpenTelemetry spans for runs (today), or per step, model and tool call (planned); the OTLP exporter is optional.
• Logs: structured JSON with the run ID on every line, with secrets redacted.
• Health: /healthz (process is alive) and /readyz (database and queue reachable).

Scrape the metrics with Prometheus or any compatible system; the Helm chart ships a ServiceMonitor for the Prometheus Operator.