add hostIP as env to server-statefulset and substitute env variables in extraConfig

[kschoche]

Changes proposed in this PR:

• Add hostIP as env variable to server statefulset
• Copies the extraConfig configMap file to /consul/extra-config/extra-from-values.json and uses sed to replace HOST_IP/POD_IP/HOSTNAME at runtime with the environment.
• Passes -config-file=/consul/extra-config/extra-from-values.json to server-statefulset and client-daemonset.

Resolves #657.

How I've tested this PR:
unit tests updated and pass, manually tested using the extra-env-args telemetry specified in issue #657 :

global:
  imageK8S: kschoche/consul-k8s-dev
  image: hashicorp/consul:1.10.0
  tls:
    enabled: true
  acls:
    manageSystemACLs: true
  metrics:
    enabled: true
controller:
  enabled: true
server:
  extraConfig: '{ "telemetry": [ { "disable_hostname": true, "dogstatsd_addr": "HOST_IP:8125", "dogstatsd_tags": ["env:weweq","service:consul","version:1.10.0"] } ] }'
  replicas: 1

And confirmed that the server and clients comes online with the changes and fails to come online with existing code.

How I expect reviewers to test this PR:
unit tests pass, code review.

Note to reviewers:

1. Given this is not configurable and just becomes hard-coded text in server-statefulset and client-daemonset is there anything worth additionally testing in bats?
2. In the current design consul will parse the config-file twice, once because it's in the configMap file /consul/config/extra-from-values.json and once again in the new one. Based on https://www.consul.io/docs/agent/options#_config_file merging should be "ok". I'm not sure this is guaranteed. If we are concerned about that, I can rename the configMap entry to extra-from-values.tmp and Consul will ignore it when processing /consul/config/ and only catch the one we modify. Thoughts?

Checklist:

• Bats tests added
• CHANGELOG entry added (HashiCorp engineers only, community PRs should not add a changelog entry)

[thisisnotashwin]

I looked at the comment regarding the testing. I think i agree with the complexity of it. I havent tested this PR, but if you have and it does solve the problem, even when the input json is complicated, this LGMT!!

Thanks for putting in the hard work on this one. It definitely was quite a wild journey for a PR!!

[ishustava]

Looks good!