mirror of
https://github.com/k3s-io/kubernetes.git
synced 2026-07-18 04:54:54 +00:00
6b4a37194a
Currently, we get each released PV every 15s, and in parallel. If there are a lot of released PV and we cannot finish all the get in 15s, it will starve other request by making the queue waiting for client-side throttling very long. Even in a normal cluster, these requests are taking majority of all get requests from KCM (58% or 470 qps) in our stress test.