I'm running a Ruby on Rails application against Postgres servers. For a long time, I've struggled with the background delayed_job process falling over at seemingly random times, and needing to be restarted. The error messages in the delayed_job stack traces are ambiguous, but they seem to suggest that the database instance is being interrupted.
The other morning, delayed_job and three other, long-running jobs were all killed. Thanks to networking problems at my company, I've finally started logging on my database instance, and it looks like these process-killing problems are happening when the database logs an error code of "08P01". According to PSQL docs, that's a connection exception with a protocol error. Nothing else should be connecting to the database right now, while the "housekeeping" is going on.
Maybe is could be the "critical" updates that it says will be automatically applied? I can't imagine that these are happening around once a week, and I can't find a log of this activity. Seems like there should be a public list to check against. Maybe these are logged, and now that I'm logging, I'll see that?
This feels like some sort of housekeeping that Microsoft is doing on Azure, restarting the database, or rerouting traffic to a different backhaul, but I don't know what that would be, let alone how to stop it (or at least schedule it). Does anyone know what could be interrupting my database connections every "several" days, and killing my jobs?
2022-12-09 11:16:15 UTC-639318ff.af6-LOG: invalid length of startup packet 08P01
2022-12-06 22:23:51 UTC-638fc0f7.5eb-LOG: invalid length of startup packet 08P01
I now see that I have 2 of these on the 8th. This is what I expected to get for "security" updates, but didn't know where to look. Unfortunately, it doesn't correlate with errors from the 6th and the 9th.
At Thursday, December 8, 2022 at 6:38:57 AM EST, the Azure monitoring system received the following information regarding your Azure Database for PostgreSQL Flexible Server:
Due to maintenance task being performed on your Azure Database for PostgreSQL - Flexible server, the server is temporarily unavailable.
At Thursday, December 8, 2022 at 6:38:49 AM EST, the Azure monitoring system received the following information regarding your Azure Database for PostgreSQL Flexible Server:
Due to maintenance task being performed on your Azure Database for PostgreSQL - Flexible server, the server is temporarily unavailable.
