{"slug":"implementation/observability-operations","title":"30 · Observability & Operations","section":"implementation","sectionLabel":"Implementation","status":"unclassified","type":"document","path":"implementation/observability-operations.mdx","bytes":1648,"description":"Execution traces, Work state, agent state, pipeline monitoring, reconciliation, failures, retries, health, and the operational console.","updated":null,"source":"integratewise/Integratewise-docs-wiki","content":"\n## What Operators Watch\n\n* **Execution traces** — Work → plan → governance verdict → Hermes execution → evidence → verification → reconciliation → Spine.\n* **Work state** — every Work's current state and history, including handoffs and evidence.\n* **Agent state** — Spine Agent identity, continuity, and Work ownership.\n* **Pipeline monitoring** — four flows, loader/normalizer health, governed-writer lag.\n* **Verification & reconciliation** — expected vs observed rates, mismatch classes, retry/compensate outcomes.\n* **Health** — tenancy isolation, projection invalidation, and rehydration latency.\n\nFailures are **expected** (mismatch, partial success, stale reads) and have first-class reconciling paths — see [Execution Closure](/closure/execution-closure).\n\n## Operational Console\n\nThe console surfaces Work, execution, pipeline, and audit — tenant-scoped, identity-bound, governed.\n\n## Where to Go Next\n\n<CardGroup cols={2}>\n  \n    How outcomes become truth or get reconciled.\n  \n  \n    The runtime being observed.\n  \n  \n    How verification is tested before it runs in production.\n  \n</CardGroup>\n"}