Update config, remove deployment, add revoke cronjob

[anna-parker]

resolves #1777

preview URL: https://ingest-deployment.loculus.org/

Summary

• We currently sometime times have dual ingest due to the scheduled cronjobs launching at the same time as the post-launch deployment. Fix this by removing the deployment and only having cronjobs that try to start every 2 minutes (deleted if cannot start up in 1 minute) but are not allowed to be concurrent and will run for max 30min.
• Additionally add a cronjob that contains revoked. Triggered by: f6bc4a4#r145569186 - getting a rule to run in a container posthoc is hard, instead just create a cronjob that never runs and launch it manually when we are ok with the suggested revocations.

Screenshot

PR Checklist

• Make sure concurrency is working correctly on argocd: This works perferctly it will just mean that the jobs are always deleted and we see this:
  - hope it is not an issue
• Make sure revoke cronjob can be started manually, using kubectl create job --from=cronjob/loculus-revoke-and-regroup-cronjob-{config.organism} <manual-job-name>
  Start works:
  , then it also shows up in argocd. And it is deleted after 30min as desired:

[anna-parker]

I had to add the approve step or jobs were starting very quickly after submission and I could still see concurrency - adding the approve step ensures jobs will run for the full 30minutes.

[corneliusroemer]

Nice! What about making approve quit by itself after 25 minutes? Then we get happy ending and not the kill from kubernetes.

[anna-parker]

Sadly even with argocd.argoproj.io/sync-options: Force=true,Replace=true the jobs are not being replaced on sync - I am not sure what is going on here

Update: extensive testing in #2478 (comment) has shown that argocd is actually restarting jobs on resync this just isn't displayed that well