[chart/redis-ha][BUG] redhat w/ ulimit unlimited max out cpu & memory, could we add to helm chart a default maxconn?

[lknite]

Describe the bug
On some os's ulimit ends up being unlimited & haproxy will max out cpu & memory in this case.
Reference: docker-library/haproxy#194

To Reproduce
haproxy/haproxy#2121

Steps to reproduce the behavior:

1. Deploy kubernetes on redhat 9
2. Deploy redis-ha helm chart

Expected behavior
haproxy to deploy normally

Additional context
Seems a good idea to specify a reasonable value in the haproxy helm chart:

[lknite]

I'm using an argocd helm chart which uses this helm chart as a subchart, if maxconn were available to be set I could configure this to solve the issue (i think).

Since this seems to be a known issue with the solution being to set maxconn, could we get this added to the helm chart as a value which can be set? Doing so would seemingly avoid an insta-crash on yum-based systems.

Or is there a way to set it via the values.yaml I'm just not seeing?

[lknite]

I worked around this by adjusting ulimits at the container service level, in my case this is containerd fixed like this:

# sed -i 's/LimitNOFILE=infinity/LimitNOFILE=65535/' /usr/lib/systemd/system/containerd.service
# systemctl daemon-reload
# systemctl restart containerd
# k delete deployment <asdf>

[wangycc]

I also encountered the same problem in the Linux 6.x kernel. Why is the same LimitNOFILE configuration working fine in lower versions?