Disable resource requests in Helm CI runs

[dimitarvdimitrov]

Signed-off-by: Dimitar Dimitrov dimitar.dimitrov@grafana.com

What this PR does

During the development of #2087 helm CI runs (call-lint-test / lint-test) started failing because pods could not be scheduled (example). The info at the end of the run which describes all pods showed the following for two pods:

Events:
  Type     Reason            Age                 From               Message
  ----     ------            ----                ----               -------
  Warning  FailedScheduling  11s (x13 over 10m)  default-scheduler  0/1 nodes are available: 1 Insufficient cpu.

GitHub hosted runners only have 2 CPUs and 7GB of memory (docs).

This PR removes resource limits and requirements for some pods in the CI values files for Helm. This way Kubernetes won't have to calculate resource quotas. The caveat is that now CI runs may become less stable because of resource contention.

The changes here were cherry-picked onto the work in #2087 to demonstrate that they unblock the CI runs - example run

An alternative approach is to install self-hosted runners with more available CPU and memory. A second alternative is to migrate the CI runs to drone, for which we have bigger agents.