Skip to content

Workaround for 503s during scaling down should not involve changing Istio sidecar injector #1370

@tcnghia

Description

@tcnghia

Expected Behavior

When scaling down, we don't see requests hitting downed Pods causing 503s. Workarounds for this issue shouldn't include changing Istio sidecar injector either.

Actual Behavior

Previously we see this behavior. We worked around by adding a PreStop sleep to the sidecar proxy, so that the containers stay around a little bit longer.

However this workaround isn't robust if the users have Istio settings of their own. We need to fix this.

Metadata

Metadata

Assignees

Labels

area/autoscalearea/networkingkind/bugCategorizes issue or PR as related to a bug.kind/cleanupCategorizes issue or PR as related to cleaning up code, process, or technical debt.kind/docSomething isn't clear

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions