Memory usage for Contour doubled in latest version

[Neekoy]

What steps did you take and what happened: [A clear and concise description of what the bug is.]

We upgraded from 1.23.2 to 1.31.0 on one of our environments, and suddenly the new Contour/Envoy containers started getting killed due to OOM exception.

We increased the resources available for the pods, which mitigated the issue.

Looking at the new and old deployments, the new ones are using twice as much RAM.

What did you expect to happen:

The new version to not use double the RAM of the old version.

Anything else you would like to add: [Miscellaneous information that will assist in solving the issue.]

This is the new version:

kubectl get pod contour-contour-695d899d54-gg94h -n contour -o jsonpath='{.status.containerStatuses[*].image}'
docker.io/bitnami/contour:1.31.0-debian-12-r

k top pod -n contour
NAME                               CPU(cores)   MEMORY(bytes)
contour-contour-695d899d54-gg94h   3m           203Mi
contour-contour-695d899d54-jphv6   1m           1582Mi
contour-contour-695d899d54-xt6k4   3m           205Mi
contour-envoy-2n7tb                6m           905Mi
contour-envoy-4jhn2                4m           895Mi
contour-envoy-5jv8c                5m           913Mi
contour-envoy-5rvfb                5m           894Mi
contour-envoy-6m2hg                4m           903Mi
contour-envoy-dtn4q                5m           899Mi
contour-envoy-htwvw                4m           892Mi
contour-envoy-jpcnn                5m           904Mi
contour-envoy-k9mfq                4m           900Mi
contour-envoy-s49pg                5m           896Mi
contour-envoy-vxxll                5m           902Mi
contour-envoy-xb67f                5m           908Mi
contour-envoy-zx6zw                5m           889Mi

And this is the old version:

kubectl get pod contour-contour-f6cc6cc7c-w6d29 -n contour -o jsonpath='{.status.containerStatuses[*].image}'
728080024150.dkr.ecr.eu-central-1.amazonaws.com/docker-hub/bitnami/contour:1.23.2-debian-11-r0

kubectl top pods -n contour
NAME                              CPU(cores)   MEMORY(bytes)
contour-contour-f6cc6cc7c-w6d29   30m          972Mi
contour-envoy-2wxsh               17m          379Mi
contour-envoy-5scjg               35m          394Mi
contour-envoy-6mckd               28m          381Mi
contour-envoy-8c6k5               33m          384Mi
contour-envoy-fl77r               21m          385Mi
contour-envoy-gwcfc               25m          385Mi
contour-envoy-hjmtn               47m          385Mi
contour-envoy-ljqhq               30m          380Mi
contour-envoy-mkqfg               20m          380Mi
contour-envoy-p95fl               31m          380Mi
contour-envoy-qnwvz               33m          384Mi
contour-envoy-srwql               30m          385Mi
contour-envoy-sxdmg               23m          384Mi
contour-envoy-xhbxd               30m          384Mi
contour-envoy-zjs9v               29m          381Mi
contour-envoy-zlwk6               22m          387Mi

As you can see the RAM usage has more than doubled, which is rather concerning. Environment:

• Contour version: 1.31.0
• Kubernetes version: (use kubectl version): v1.31.7-eks-bcf3d70
• Kubernetes installer & version: ArgoCD v2.8.0
• Cloud provider or hardware configuration: AWS EKS
• OS (e.g. from /etc/os-release):