Open mbrancato opened 2 years ago
This issue is stale because it has been open for 90 days with no
activity. It will automatically close after 30 more days of
inactivity. Reopen the issue with /reopen
. Mark the issue as
fresh by adding the comment /remove-lifecycle stale
.
/lifecycle frozen
Adding to this, I identified that the previous revisions will fail to scale down completely if listening on localhost.
Describe the feature
Currently, all liveness probes are sent to the service container, which is expected to be listening on
0.0.0.0
on the container port. There is no reason that the service container needs to listen on all interfaces if all traffic is to be proxied viaqueue-proxy
anyway. It would be good to support services that listen onlocalhost
and to whichqueue-proxy
will proxy the requests.Today, the following config will send all liveness probes to the container, and readiness probes seem to be ignore or handled by the queue proxy (not sure, might be a separate issue).
It would be nice if we didn't have to expose the service on the pod network for this to work by allowing the
queue-proxy
to handle both liveness and readiness.e.g.