[Dynamic infra] Configure pods priority

We should start listing the important pods running on the archive cluster and assign (sort) them by priority.

It will allow to assign them a priority class to ensure the most important node will always run.

For example, recently, cassandra replayers were just declared on the staging cluster and prevented the cvs loader from starting due to a lack cpu resources.

The worker pods should preempt the replayers resources.

It should be implemented in the helm chart and generalized to all the deployments.

Priority class document iteration [1].

Plan:

• Analyse workloads
• Define a first iteration of priorities [1]
• swh/infra/ci-cd/swh-charts#140: Adapt swh (& software-stories) charts to allow priority configuration
• (review) swh/infra/ci-cd/swh-charts!141 (merged): Configure cluster-components chart with priorities
• (review) swh/infra/ci-cd/swh-charts!141 (merged): Unify the way priorities are configured
• Deploy on staging
  • software-stories deployment
  • swh deployments
  • cluster-components deployments
• Deploy pod priority in production
• Status ok