RKE2-kube-apiservers consumes too much memory. Can you help me? #6493
Replies: 1 comment 5 replies
-
|
What critera are you using to determine that this is too much memory? The apiserver static pods run without memory limits by default, so they will attempt to use as much memory as is necessary to maintain a stable working set and cache. This is directly driven by client load - the more resources in the cluster, and the more client requests, the more resources the apiserver will consume. You could try putting memory limits on the apiserver pods, but this will likely negatively affect performance. |
Beta Was this translation helpful? Give feedback.
-
|
@brandond Shortly after sharing this post, server-1 crashed due to a reason that I think was caused by kube-api-server. |
Beta Was this translation helpful? Give feedback.
-
|
What makes you think that? The rke2 apiserver is unmodified from upstream Kubernetes. Whatever is going on is unlikely to be unique to rke2. |
Beta Was this translation helpful? Give feedback.
-
|
@brandond Thank you for your return. |
Beta Was this translation helpful? Give feedback.
-
|
I am not aware of any memory leak in the apiserver. If you don't put a limit on it, it will use as much memory as it considers necessary. What is "normal" depends heavily on what you're doing with the cluster. |
Beta Was this translation helpful? Give feedback.
-
|
Thank you very much, maybe because ofthis audit policy, it consumes a lit bit more memory. |
Beta Was this translation helpful? Give feedback.
-
Log for kube-apiserver and I am using Kubernetes Version: v1.30.0 +rke2r1
This is config yaml for server.
Beta Was this translation helpful? Give feedback.
All reactions