pmo: round budget 30, refeed 522/495 w/ WIP discipline, file 526 writeback retry
This commit is contained in:
@@ -0,0 +1 @@
|
||||
Three Cloudron flaps today (11:28Z 502, 13:45Z 502, 14:11Z 500). Loop scan errors retry naturally, but status/note writebacks fail permanently when they land in a flap (case 519 at 11:28Z: work done, all three writes lost, ticket left New — PMO repaired manually; a PMO batch at 12:50Z was similarly swallowed while reporting success). Add bounded retry with backoff + verification re-read for status/note writes; on final failure park the task (leave New + local flag) and surface last_writeback_error in the heartbeat status.json. Filed by PMO after audit.
|
||||
Reference in New Issue
Block a user