Increase Mutating webhook timeout.

[alpeb]

Discussed in https://github.com/linkerd/linkerd2/discussions/11984

^{Originally posted by smithdebug January 24, 2024} Hello Team, I’m reaching out to inquire about the possibility of extending the timeout for the linkerd-proxy-injector-webhook-config Mutating Webhook. Currently, the timeout is set to 10 seconds bydefault. We’ve recently faced a situation where this timeout caused an issue, resulting in a Pod without a Linkerd proxy sidecar. Thanks

[alpeb]

@smithdebug , would you be willing to tackle this issue? Sounds like a relatively simple setting to add and hook in the helm templates.

[alpeb]

Thanks for opening #12028, we'll take a look at it soon. OTOH, we're still interested on why you actually bumped into this issue. The proxy injector should not take that long to process pods. Can you share a bit more about your setup? Is there a lot of pod churn, maybe during node escalations, causing requests to the injector to spike? Are you already running in HA mode, increasing the number of the injector replicas?

[smithdebug]

We have around 3000 pods and 5 replicas for proxy-injector. We have enabled webhook failure mode, earlier it was set to ignore.

[wim-de-groot]

Is this issue still relevant?

[mateiidavid]

@wim-de-groot nope! thanks for pointing it out.

Closing, since this has been resolved through https://github.com/linkerd/linkerd2/pull/12028. Thanks all!