Higher idle process activity after upgrade

Cool hint. I didn’t see anything alarming there, no jobs running and none with long execution (only a few over 100ms). Dashboard says ~1000 processed per day, same as before the upgrade.

Probably off-topic, but interesting also the Processed graph over 6 months, seems like there are two plateaus (one twice as high as the other) that it toggles between with weeks between. It does not correlate with reboots (and we didn’t update this stuff in that time either).

Is there an easy way to change that 2s BRPOP timeout? I didn’t find it in the code. It would at least indicate whether the work polling loops are the problem.