really at quite a loss as to what to keep troubleshooting. the worker node is BEEFY, definetely not a resource limit. I've increased the control plane resources to like 16 cores and 64GB, plenty of disk space. no network issues that I can see