Release test serve_autoscaling_multi_deployment.aws failed

[can-anyscale]

Release test serve_autoscaling_multi_deployment.aws failed. See https://buildkite.com/ray-project/release/builds/6224#018d2106-5fbb-41de-90fd-f597426a43f2 for more details.

Managed by OSS Test Policy

[can-anyscale]

Test has been failing for far too long. Jailing.

[anyscalesam]

@zcin do you have bandwidth this week to look into this? cc @akshay-anyscale

[zcin]

@anyscalesam I took a look at this and it seems to be failing for some obscure reason. Will take a bit more time to look into it, but may leave it jailed for now and rewrite it since we are rewriting a good number of our release tests for 2.10.

[anyscalesam]

@zcin any ETA on the release tests rewrite?

[edoakes]

@anyscalesam this was pushed back due to Cindy prioritizing draining/compaction improvements instead. Release tests won't get reworked until 2.11.

This test failure shouldn't be a blocker for 2.10, however (it's archaic).

[anyscalesam]

Thanks @edoakes; how should we think about this test in with respect to getting to a weekly green build from master? Is it important enough that it needs to stay within that scope?

[edoakes]

Thanks @edoakes; how should we think about this test in with respect to getting to a weekly green build from master? Is it important enough that it needs to stay within that scope?

No, as written it's not, we can even just disable this test.

[can-anyscale]

I'll put up a PR to mark this test as unstable to keep the record ;)

[can-anyscale]

oops, I mistakenly pushed the change directly to master https://github.com/ray-project/ray/commit/c6094a96aa9e08b43cfdd15a7c2a1d07eddd2dad

[can-anyscale]

Re-opening issue as test is still failing. Latest run: https://buildkite.com/ray-project/release/builds/10617#018e1301-5874-4c5c-9ebe-e24026405245

[zcin]

This has been removed from our release tests and replaced with serve_autoscaling_load_test