Function domain is hardcoded as .svc.cluster.local and this is not correct for all clusters

[Jc2k]

Is this a BUG REPORT or FEATURE REQUEST?:

Bug report

What happened:

When generating a request to a function it assumes the domain name is .svc.cluster.local. See here.

What you expected to happen:

One of:

• The suffix is just .svc - the dns service will automatically expand it to the full form
• nats-trigger automatically works out the correct FQDN
• there is a config option to set it.

How to reproduce it (as minimally and precisely as possible):

I have a kubeadm cluster where the networking/dnsDomain setting of ClusterConfigration was set to a domain, not cluster.local.

Anything else we need to know?:

Environment:

• Kubernetes version (use kubectl version): 1.16
• Kubeless version (use kubeless version): 1.0.6
• Cloud provider or physical cluster: bare metal

[jdmeyer3]

I was able to get around this issue by editing the yaml file like so

---
apiVersion: apps/v1
kind: Deployment
metadata:
  labels:
    kubeless: nats-trigger-controller
  name: nats-trigger-controller
  namespace: kubeless
spec:
  selector:
    matchLabels:
      kubeless: nats-trigger-controller
  template:
    metadata:
      labels:
        kubeless: nats-trigger-controller
    spec:
      containers:  
      - env:
        - name: "NATS_URL"            # domain override here
          value: "nats://nats.<namespace>.<custom domain>:4222"

adding the "NATS_URL" env value allowed me override the default cluster domain

[jdmeyer3]

A quick correction, the trigger controller comes online with my workaround. However, the trigger is still sending requests to "svc.cluster.local" domains. So a fix would still be needed