docs: route queue heartbeat evidence gap
All checks were successful
CI Smoke / host-smoke (push) Successful in 0s
CI Smoke / container-smoke (push) Successful in 2s

Assistant: codex
Assistant-Model: gpt-5.6-sol
Assistant-Session: 01a06bfe-2a55-7ed3-bacd-879977b099bf
This commit is contained in:
tegwick 2026-09-04 20:14:26 +02:00
parent df34afd26a
commit fc38ca0fdf

View file

@ -9,7 +9,7 @@ owner: codex
topic_slug: activity-core
priority: high
created: "2026-08-23"
updated: "2026-08-23"
updated: "2026-09-04"
related:
- ACT-ADR-005
- ACTIVITY-WP-0026
@ -149,3 +149,14 @@ No run was claimed during the observation window, so a natural active
heartbeat and rejected late close remain before T04 can be marked done. The
sole open row was an intentionally non-automated retention record and was not
modified to manufacture evidence.
Follow-up 2026-09-04: 24 natural runs have been claimed since rollout and all
terminated in 534 seconds. The rein consumer schedules its first heartbeat at
one third of the 900-second lease, capped at 300 seconds, so production API
logs contain accepted claim/fail traffic but zero heartbeat requests. This is
an observability gap in the consumer timing, not evidence that Activity Core's
lease enforcement is absent. State Hub message
`1250c36a-9fe8-45dc-bb5a-d5fa85e49fd0` asks rein-aharness to heartbeat once
immediately after claim, retain periodic heartbeats, and return a natural
claim/heartbeat/close trace. It also requests an isolated labeled late-close
smoke; no real workload will be delayed or expired to manufacture evidence.