Closed jackleland closed 9 months ago
Root cause fixed in 437ad571b1b8395cbabfe82d64a3e125b2c77f33
Initial cause/discovery of bug was the archive-put job not having a default name to put into the monitoring db, fixed as of ad487f43391dfe7676b713e744f18b10852dcec4
Will close when merged into main.
The generic failure mode for a consumer is to remove the message from the current queue and forward something to the monitor to alert the monitoring db to the failure. This will create an infinite loop if the monitor is the place of failure as the message is immediately requeued to the monitor. Either a read of the retries needs to happen or the monitor needs to not forward the message on to itself, preferably both.