In my case I started to get this error when I wanted to check the logs of my Next.js app in a EKS. The app page were loading but when the app API backend wanted to contact an external API it was failing, returning 502. I discovered that the NAT gateway for my cluster subnet was changed and the previous on was not available any longer. I had to change the route table for the subnets to point to the new NAT gateway. Once this was done, everything started working.

The reason for the error was a VPC network configuration change were the connection to external systems in the cluster was not working therefore causing fetching the logs from my PC to not be able to reach the pod even though I could list, get or describe Kubernetes resources.