Unable to create a fluentdconfig with an environment variable replacement

[allenmun197]

Describe the bug

This bug seems to come from the reloader as creating a fluentdconfig with an env var replacement like "#{ENV['HEC_ENDPOINT']}"

When deploying a fluentdconfig like

apiVersion: logs.vdp.vmware.com/v1beta1
kind: FluentdConfig
metadata:
  name: fluentd-config
spec:
  fluentconf: |-
    <match **>
        @type splunk_hec
        format json
        hec_host "#{ENV['HEC_ENDPOINT']}"
        hec_port 8089
        hec_token <hec_token>
        index some_index
        sourcetype app_log
        reconnect_on_error true
        reload_on_failure true
        reload_connections false
      <buffer>
        @type file
        path /var/log/buffer/
        flush_thread_count 8
        flush_mode interval
        flush_interval 10s
        chunk_limit_size 16MB
        retry_wait 5
        retry_max_times 10
        retry_forever false
        retry_timeout 2h
        retry_exponential_backoff_base 2
      </buffer>
    </match>

It will error out the validator in the reloader. Seeing this in our logs.

time="2022-04-14T14:55:33Z" level=debug msg="Checked config for namespace default with fluentd and got: 2022-04-14 14:55:33 +0000 [error]: config error file=\"/tmp/validate-ext-default4277819843\" error_class=Fluent::ConfigError error=\"One of `hec_host` or `full_url` is required.\""
time="2022-04-14T14:55:33Z" level=error msg="error running validation command: exit status 2"
time="2022-04-14T14:55:33Z" level=info msg="Configuration for namespace default cannot be validated with fluentd validator"

This did not seem to be a problem with v1.14.1 of the kube-fluentd-operator. Currently we are using v1.16.3.

Reproduction steps

1.  Install kube-fluentd-operator from the in repository helm-chart
2.  Apply the given manifest

apiVersion: logs.vdp.vmware.com/v1beta1
kind: FluentdConfig
metadata:
  name: fluentd-config
spec:
  fluentconf: |-
    <match **>
        @type splunk_hec
        format json
        hec_host "#{ENV['HEC_ENDPOINT']}"
        hec_port 8089
        hec_token <hec_token>
        index some_index
        sourcetype app_log
        reconnect_on_error true
        reload_on_failure true
        reload_connections false
      <buffer>
        @type file
        path /var/log/buffer/
        flush_thread_count 8
        flush_mode interval
        flush_interval 10s
        chunk_limit_size 16MB
        retry_wait 5
        retry_max_times 10
        retry_forever false
        retry_timeout 2h
        retry_exponential_backoff_base 2
      </buffer>
    </match>

3. Check the logs on any log-router pod in the reloader container

### Expected behavior

The reloader should be able to accept this valid fluentd config syntax and reload the config.

### Additional context

_No response_

[cneepa]

Hi, we are facing the same issue after upgrading from 1.15.1 to 1.16.7.
What will be the version which might have introduced the bug?