Drive apps are deployed as statefulsets with poddisruptionbudgets of maxUnavailable=0 to prevent intermittent downtime. But this prevents node draining in case it's needed.
Ideas: a job that temporarily updates the PDBs to allow disruption? Or a PDB of some kind that allows for draining but not disruption for cluster scale-down (maybe this is easy/default?)
Drive apps are deployed as statefulsets with poddisruptionbudgets of maxUnavailable=0 to prevent intermittent downtime. But this prevents node draining in case it's needed.
Ideas: a job that temporarily updates the PDBs to allow disruption? Or a PDB of some kind that allows for draining but not disruption for cluster scale-down (maybe this is easy/default?)