Does Kamal's `drain_timeout` work with SolidQueue?

[brenogazzola]

I've noticed some of my jobs that take 10-20 seconds to complete are getting killed during deploys, which is weird, considering Kamal's default drain timeout is 30.

A quick search here makes no mention of drain_timeout, so I'd like to confirm if the problem is the lack of support for solid_queue, or something else in my configuration.

[rosa]

Oh! @brenogazzola, yes, this should work just fine with Solid Queue. There's no direct support for that timeout, but the support is based on the different signals sent to the processes to tell them they should initiate a shutdown and not pick up more jobs. I wonder if there's been a change from Kamal 1 to 2 that has affected the signals sent or something else, though... @djmb, do you know off the top of your head?

[djmb]

No there shouldn't be any changes. We shut down the container with docker stop -t <drain_timeout>, so docker will send a SIGTERM immediately and then a SIGKILL after 30 seconds, if using the default value.

The only thing I can think of that might change that is if your Dockerfile overwrote the STOPSIGNAL, or there were options in the deploy config to set a different stop signal when booting the container.