I have a long-running database migration that I intend to run through Garden. I've defined it as a Run Action with a timeout time of 100 hours (360,000 seconds), but I'm seeing it fail after almost exactly 45 minutes very consistently. The message I receive is:

⚠ run.migrate-db     → Failed to finish streaming logs: Timed out after 20 seconds. Some logs might be missing in the verbose log output, or in Garden Cloud.
✖ run.migrate-db     → Failed (took 2793.3 sec)
✖ run.migrate-db     → Failed processing Run type=kubernetes-pod name=migrate-db (took 2796.41 sec). This is what happened:

────────────────────────────────────────────────────────────────────────────────────────────────────
Failed to start Pod run-migrate-db-090f8e.

Here are the logs until the error occurred:

[Could not retrieve previous logs for deleted pod run-migrate-db-090f8e: Error: Error while performing Kubernetes API operation readNamespacedPodLog: ApiException

Response status code: 400
Kubernetes Message: previous terminated container "main" in pod "run-migrate-db-090f8e" not found]
────────────────────────────────────────────────────────────────────────────────────────────────────
1 requested run action(s) failed!

See .garden/error.log for detailed error message

As far as I can tell, this looks like the Pod was intentionally killed, assumably by Garden since there are no other external processes managing pod lifecycles. Is there some internal maximum that overrides the

timeout

value of an action?