Pick an option on each card and press Done; your answer goes straight
back to the session. The raiser's pick is marked. "I'm not sure" is a real answer.
1
In plain terms. The background worker that handles renewals, maintenance and tours is running month-old code after someone rolled it back by hand tonight, which brought back an old, already-fixed bug — should I put it back on today's code, or is the rollback intentional?
The full question, as raised
propflow-renewal-worker-prod's ECS service is running task-def revision 4328 (image 9f9682063bcc, commit 2026-09-05), registered directly by IAM user fede-admin at 2026-09-30T20:42:13-05:00 — overriding revision 4327 (image 61877475d096, today's legitimate CI deploy at 17:57:11-05:00) which already carries both fixes for this exact alert (PR #7944: stop minting one Sentry issue per household; PR #7971: skip sandbox/not-live orgs). The reverted window also drops ~2 months / 5273 commits of other merged fixes across every Temporal queue this worker services (renewal, tour, turnover, maintenance, leasing, collections, etc). Should the service be redeployed to current main (or pinned to 4327), or does the Sept-5 pin need to stay for a reason I can't see from here?
Why, and what was tried
The session raising this recommends: Redeploy to today's code The rollback predates two already-shipped fixes for this exact alert and silently reverts ~2 months of merged work platform-wide; I found no comment or ticket explaining it, so I'm not undoing a human's very recent hand action on production without confirmation. Tried first, unsuccessfully: Pulled the Sentry event's raw breadcrumbs (context=PmActionReminderWorkflow), matched the message to src/lib/temporal/activities/pm-action-reminder.ts, confirmed the org (Fairhaven Residential TEST, isSandbox=true) should be silenced to a warn per the code on main. Traced the deployed ECS task definition and image tag, diffed it against origin/main, and found the currently-running image is from 2026-09-05 — 5273 commits behind — registered by a human IAM user minutes before the alert fired, overwriting a same-day automated deploy that already had the fix. Stopped short of redeploying myself since that would directly reverse a production change Fede appears to have just made on purpose.