To be able to dynamically wait that no more connexion exist on varnish (app level) before to terminate a pod

To check if "MEMPOOL.sess[0-9]+.live" stats are relevant for all cases.

#8

https://artifacthub.io/packages/helm/varnish/varnish-cache Currently does not have verified badge

• Don't provide server.vclConfig
• When vclConfig is empty, use the default VCL in the Docker container.

Please provide TLS support by adding an extra sidecar in the pod for Hitch.

• Hitch has an official Docker image and easily be pulled in
• The container can expose itself over port 8443 and the service can expose that over 443
• The --backend parameter can be used to connect to Varnish, possibly over UDS instead of TCP
• The --frontend parameter can be used to listen on port 8443 and to set the location of the TLS certificate
• The --write-proxy-v2 flag can be enabled to communicate with Varnish using the PROXY protocol. This only works if there's an interface on Varnish that listens to PROXY traffic
• The --alpn-protos option can be used to offer HTTP/2 support

Please also make sure the way to enable TLS matches the syntax of the Varnish Enterprise Helm Chart. This means providing the following Helm config overrides:

• server.tls.enabled
• server.tls.port
• server.tls.config
• server.service.https.enabled
• server.service.https.port
• server.service.https.nodePort

Thanks for considering.

Currently, script/get-latest-tag.sh is retrieving alpine image instead of non-alpine ones (as it's based on version sorting, and suffixed image takes priority). We should change the script so that the non-suffix image takes the priority (i.e., 7.5.0 instead of 7.5.0-alpine).

To be able to customize the behavior once hpa enabled

cc #6

server.extraEnvs is missing (for some reason)

I wanted to use the official Varnish Helm chart to deploy a Varnish cluster with HPA enabled.

When I send a PURGE request, Can you explain how purging works in this setup and provide recommendations for ensuring consistent purging across all replicas?

Wanted to ensure cache invalidated across all replicas & get the acknowledgment for the same.