test installation gives 502 Bad Gatewayon (kubespray provisioned on KVM vms)

Summary

Getting a 502 Bad Gateway error trying to access https://gitlab.example.com/, which I set up in my local /etc/hosts file to point to the externalIP I specified in the config.

Steps to reproduce

Stand up new cluster using Kubespray, which uses MetalLB by default.

Configuration used

values:
  global:
    ingress:
      configureCertmanager: false
      #class: "nginx"
    hosts:
      domain: example.com
      externalIP: 192.168.96.131
      https: false
    rails:
      bootsnap:
        enabled: false

  #  loadBalancerIP: 192.168.96.130
  # --set certmanager-issuer.email=me@example.com
  gitlab:
    webservice:
      minReplicas: 1
      maxReplicas: 1
      resources:
        limits:
         memory: .5G
        requests:
          cpu: 100m
          memory: 100M

  gitlab-runner:
    install: false
  gitlab-shell:
    minReplicas: 1
    maxReplicas: 1
  prometheus:
    install: false
  registry:
    hpa:
      minReplicas: 1
      maxReplicas: 1
  # Don't use certmanager, we'll self-sign
  certmanager:
    install: false

Current behavior

Go to https://gitlab.example.com/, get a 502 Bad Gateway error.

Expected behavior

Initial setup screen?

Versions

  • Chart: gitlab/gitlab (so latest?)
  • Platform:
    • Self-hosted: (local Kubespray based cluster running on KVM )
  • Kubernetes: (kubectl version)
    • Client: v1.22.3
    • Server: v1.21.3
  • Helm: (helm version)
    • Client: 3.7.0
    • Server: 3.7.0

Relevant logs

(Please provide any relevate log snippets you have collected, using code blocks (```) to format) kubectl get pod shows

gitlab2-webservice-default-78d8d77dc9-n7x2p            1/2     CrashLoopBackOff   17         80m   10.233.68.7    kube-node03   <none>           <none>            app=webservice,chart=webservice-6.0.0,gitlab.com/webservice-name=default,heritage=Helm,pod-template-hash=78d8d77dc9,release=gitlab2

kubectl describe pod gitlab2-webservice-default-78d8d77dc9-n7x2p --namespace gitlab2 shows

Events:
  Type     Reason     Age                    From     Message
  ----     ------     ----                   ----     -------
  Warning  Unhealthy  55m (x54 over 75m)     kubelet  Readiness probe failed: Get "http://10.233.68.7:8080/-/readiness": dial tcp 10.233.68.7:8080: connect: connection refused
  Warning  BackOff    5m38s (x300 over 74m)  kubelet  Back-off restarting failed container

No logs available since the webservice container appears to not be started properly. Nothing else looks erroneous in the output of that pod, and all other pods are up and running.