GitLab

These appear to be what's causing Sidekiq to freak out and shut down all
the time, which gives us a huge latency on job processing.

Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>

Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>

note: we need to bump the registry chart, but the image doesn't
change, only the config
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>

Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>

FDO helm bot authored Jan 11, 2022 and Daniel Stone committed Jan 12, 2022

[daniels: Remove Gitaly gdb+delve additions, needs to be redone.]

Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>

gitaly-1 is doing some weird stuff, so we plan
on retiring it.
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>

Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>

minio was already disabled, and the chart itself is now returning 404.

Remove it
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>

Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>

Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>

Hopefully this will prevent us to manually kick the registry cache
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>

So we can reuse it from other namespaces without having to carry the
DNS01 secret there.
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>

Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>

registry.gitlab.com/charts/alpine-helm is too old and breaks with
let's encrypt certificates now.
Use that opportunity to actually use a recent version of alpine and
also ci-templates to not pull countless times the same packages over and
over
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>

Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>

Same fate than gitaly. Enforcing the fsgroup takes 5 min for the pod to
spin up. But only a few seconds when disabled.
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>

This one was still using the hdd+ssd storage class, meaning that any
problem on S3 was impacting redis and the whole gitlab instance.
Signed-off-by: Benjamin Tissoires <benjamin.tissoires@gmail.com>