Ingress and nginx 502 bad gateway

I can’t get Ingress working. I get 502 Bad Gateway errors from nginx. I am not running nginx in my workload, so I assume it is from the Nginx Ingress Controller.

I am running a bare-metal setup with no cloud provider with two worker nodes, one control plane, and one etcd node.

My workload is running. I can get to it through the IP address and NodePort.

Here is my Ingress YAML:

apiVersion: extensions/v1beta1
kind: Ingress
  annotations: user-gqr8f '{"":"default:","Y29wcy1tbHQvZGVmYXVsdC94aXAuaW8vL2NvcHMtbWx0LzM1MDA=":"deployment:default:cops-mlt-deployment"}' '[{"addresses":[""],"port":443,"protocol":"HTTPS","serviceName":"default:ingress-5fd39c2a15055c598e828054905973d5","ingressName":"default:cops-mlt","hostname":"","path":"/cops-mlt","allNodes":true}]' nginx / cops-mlt-service
  creationTimestamp: 2018-11-02T17:48:30Z
  generation: 8
  labels: norman
  name: cops-mlt
  namespace: default
  resourceVersion: "496522"
  selfLink: /apis/extensions/v1beta1/namespaces/default/ingresses/cops-mlt
  uid: 82ac8f76-dec7-11e8-92a0-0050569b4af5
  - host:
      - backend:
          serviceName: ingress-5fd39c2a15055c598e828054905973d5
          servicePort: 3500
        path: /cops-mlt
  - hosts:
    - ip:
    - ip:

Does anyone have any ideas on what I can do to get this Ingress working? Thanks.


Having the same issue, where path-based rules result in 502 errors, only able to reach workload using hostname and NodePort.

+1 having the same issue, setup rancher as per the startup instructions

The same issue, the Ingress load balance does not work, and got 502 bad gateway error.

The same issue (rancher 2.2, nginx 1.15.6)

I apparently am having the same issue on 2.2.3. It was working on 2.2.2.

502 means there is no working target pod to send the request to. Which usually means you have the wrong target port, or are selecting no pods.

2.2.3 contains only 2 commits to fix 1 issue which has nothing to do with ingress.

1 Like

I’ve made the mistake to enter the port of the workload instead of the exposed port of the docker container.

@vincent just because it contains 2 commits doesn’t make it work. Same problem here

I am also facing the same issue. using rancher version 2.3.3

anyone go the fix for this issue.

i use multiple ingress facing the same issue on all of them

I fell into this trap too. I figured out where I went wrong. I was using a self signed certificate in the pod. Ingress gave me this error because the self signed certificate from the pod wasn’t part of ingress’s trusted certificates. Hope this helps anyone else running into this problem