High-CPU load after upgrading to 1.25 - flanneld high CPU

charm
,
1

Hi i used Upgrading to 1.25 | Ubuntu to upgrade to 1.25. After a small pitfall with the channel configuration see: Upgrading to k8s 1.25 fails on podsecurity policy admission plugin in api-server, I managed to upgrade to version 1.25.

After some time I noticed that the CPU load of the kubernetes nodes is significantly higher than before the update. The load on the nodes even reached 99%. On the nodes you can see that flanneld has a lot of CPU load.

flanneld2
flanneld21903×193 45.5 KB

I use canal for CNI networking:

App                       Version          Status  Scale  Charm                     Channel      Rev  Exposed  Message
canal                     0.22.1+ck1/3...  active      6  canal                     stable        84  no       Flannel subnet 10.1.79.1/24

After restarting flannel on each node with:

sudo systemctl restart flannel

the issue seemed resolved, however after 5 days it appeared again and all nodes in the cluster had a load of 99%. The logs do not show any errors or irregularities.

What is wrong, any ideas? How can i debug this?

2

I noticed that the channel of the canal charm was set to stable, the Rev 84 points to 1.28/stable. Set the channel to 1.25/stable with:

juju refresh canal --channel 1.25/stable

I will comment again if the issue still persists.