Yes, and it's getting worse.
You might want to take a look at descheduler: https://github.com/kubernetes-sigs/descheduler
IIRC, it's more or less inspired by the Kingfisher's beak[1] [1] https://en.wikipedia.org/wiki/Kingfisher
That's what we did where I work. Thanks to GKE + HPA + cluster-autoscaler, our cluster grows at the same time as our req/s.
Yes, and it's getting worse.
You might want to take a look at descheduler: https://github.com/kubernetes-sigs/descheduler
IIRC, it's more or less inspired by the Kingfisher's beak[1] [1] https://en.wikipedia.org/wiki/Kingfisher
That's what we did where I work. Thanks to GKE + HPA + cluster-autoscaler, our cluster grows at the same time as our req/s.