0

I'm having a really strange issue with a particular application that's using the same Helm chart as many other apps in our cluster, and all the others are working fine.

The pods are running the application fine, I can port-forward to any pod, or the service, and see the app locally.

However when I try connecting to it from the ingress via the internet, I get a 502 error from NGINX. Even stranger, when I add liveness and readiness probes to the deployment, those are unable to connect either.

Again, this app uses the exact same Helm chart and config as many other apps that are working, and I can't find anything that's different about it between any other application. Somehow, the pods work but the cluster is unable to reach them internally. What could cause one particular deployment to act this way?

3
  • Does the app listen on loopback only? Jan 21 at 6:40
  • Also, are you using Network Policies by any chance?
    – luislhl
    Jan 21 at 11:55
  • @chupasaurus Thank you! I knew it was something that simple. Yeah all the other apps I've launched with this chart previously were listening on all IPs by default so I never noticed this issue before. This new app was set to localhost only by default. If you post an answer I'll accept. Jan 21 at 19:19

1 Answer 1

0

If the application listens on loopback only by default it's ports couldn't be accessible by other pods. port-forward works because it ignores IP addresses and kubelet makes a proxy.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.