Hello,
I migrated some of my VM from rancher 1.6 to rancher 2.x and I have some problems with the memory and furthermore with the CPU consumption. I monitor my memory usage with collectd and grafana, and I’m little bit disapointed to have a 70% memory consumption only with rancher agent (and 20% on one core CPU).
I attached the graohic showing the memory increasing over time for the light blue / orange / green machine. As an exemple, the blue machine has not been migrated and is under rancher 1.6. So I understand the changes between 1.x and 2.x with kubernetes, but how can I solved this memory and CPU issue (I suppose CPU is due to memory gc ?).
Just to let you know the green machine is an etcd/control/worker node and light blue / orange are only worker nodes. Can this be due only to logs stashing on the node ? or is it due to some misconfiguration of my nodes ?
Because on the long term I am not able to have any pods on those machines, and memory / CPU usage is only due to rancher agent.