Greetings -
Are there any performance concerns (memory/cpu) on the Rancher Server if you have a Stack (one with ~12-24 services/containers) in an active state of trying to schedule for multiple days?
Scenario: There is a worker stack that will only schedule onto hosts with label worker=true. The hosts are spot instances and only come online for ~3 days per week.
Key Question: Can I leave that stack as ‘active’ all week or would that result in some excessive overhead on the Rancher Server during the 4 days of the week in which no hosts exist that satisfy the scheduling requirement?
We currently stop the stack the majority of the week and then manually start the stack again when we know the hosts are available - that’s clearly an admin burden that I’d like to get away from.
Thanks,
Gilman