Hello, we are experiencing long time now high memory usage in the cattle-agent pods , especially in the downstream clusters with more resources. Memory spikes many times even above 20GB and cattle-agent pods are OOM killed causing service interruption to our teams using rancher to interact with the downstream clusters. Did some investigation and i found the curlpit is the

/usr/bin/agent

process and have shared my finding in !49011. Is that a known issue? Any way to tackle it?