Hey there, we had a serious incident the other day with one of our rancher server managed rke cluster , all our apiservers were burning all the hosts ram, >70GB. We spun up an extra apiserver on a very big machine, it then ate 450GB for few dozen of minutes before reducing to ~7GB. Have you seen such behavior previously ? We're running rancher 2.6.5 and 1.20.15. thanks

There was a performance related fix in 2.6.6, but I'm not sure it was ram based - https://github.com/rancher/rancher/releases/tag/v2.6.6