kubernetes-sigs/gcp-compute-persistent-disk-csi-driver

ListVolumes doesn't scale with lots of clusters in the same project

msau42 opened this issue · 5 comments

msau42 commented

If you have hundreds of clusters in the same project, each driver pod calls ListVolumes and has to filter out lot of unrelated PDs. This results in:

  1. excessive ListVolumes calls that will hit GCP API quota
  2. unnecessary processing by the driver and csi-attacher

It would be good to explore ways we can filter this list.