Skip to content

chore: try another scheduling rules for some apps

cyrinux requested to merge chore/maybe-better-scheduling into main

(scope mastodon and ttrss)

in general:

  • prefer schedule on same zone as the primary database
  • and try to spread accross the node of the same zone the pods on a same deployement
  • add some cpu resources limit, but we should add those for ALL our stuff ?
  • dont schedule mastodon sidekiq and flux updater on the same node, but, this should not be necessary if we manage well cpu resources request/limit for those?

and rename some containers name for more visibility

Edited by cyrinux

Merge request reports