We have been getting OOM killed alerts for linked-proxy for some of our pods over the last few months.
This is happening on Azure Kubernetes Service.
The diagnostics area has this message for a possible cause for the OOM killed. Was wondering if the cgroup v2 compatibility issue could be the cause?
OOM-killed pods detected and certain applications may be affected by cgroup v2 compatibility issues.
SummaryYour cluster is running Kubernetes version 1.27.7 and certain applications may be affected by cgroup v2 compatibility issues. This is a recent issue that has started affecting some earlier versions of Java, Node.js, and .NET workloads running on Kubernetes version 1.25 or greater. You may notice node memory pressure or increased memory usage for apps when compared to their memory usage before the upgrade.
These issues could be contributing to the OOM-killed pods below. Please view the resolutions below in the section titled: Cluster was recently upgraded to Kubernetes 1.25 or greater for instructions as to how to mitigate these issues.