Open
Description
We have a WAF that has been deployed to dev for quite a while, but as we move it to sandbox, we noticed that it causes the envoy admin server to start becoming unresponsive. We use our admin endpoint to serve /ready
and /server_info
for liveness and readiness healthchecks, and when this happens, Kubernetes starts killing the healthy pods.
The container doesn't seem to be resource constrained when this is happening, but I wonder if it's something like envoyproxy/envoy#16425.
We are still in the midst of debugging, but posting here in case someone has ran into this before.
Metadata
Metadata
Assignees
Labels
No labels