OOM kill on do k8s

[domino14]

this is happening like once a day:

[2023-10-30 02:03:04 +0000] [1] [CRITICAL] WORKER TIMEOUT (pid:1902781)
[2023-10-30 02:03:05 +0000] [1] [WARNING] Worker with pid 1902781 was terminated due to signal 9
[2023-10-30 02:03:05 +0000] [1960150] [INFO] Booting worker with pid: 1960150
[2023-10-30 02:03:06 +0000] [1] [CRITICAL] WORKER TIMEOUT (pid:1902782)
[2023-10-30 02:03:07 +0000] [1] [WARNING] Worker with pid 1902782 was terminated due to signal 9
[2023-10-30 02:03:07 +0000] [1960156] [INFO] Booting worker with pid: 1960156

[domino14]

I think k8s is running more crap on our cluster which is small already. We could also have a mem leak. But DO makes it difficult to see how much memory we're even using. Fargate would show this better (see #419)

Doubled memory of cluster which is more expensive.