Unable to Disable APM via the Helm Chart

[hamishdh]

Describe what happened:

I am unable to disable APM trace collection via the Helm chart. I had initially enabled it by setting the below config:

    datadog:
      apm:
        socketEnabled: true
        portEnabled: true

To disable it, I have tried setting the below values:

    datadog:
      apm:
        socketEnabled: false
        portEnabled: false

The Helm release deploys, however APM traces are still being collected. A 'trace-agent' container is still being created in the DataDog DaemonSet and it has the environment variable: DD_APM_ENABLED=true.

The 'agent' / 'process-agent' containers correctly have the environment variable DD_APM_ENABLED=false.

Also interestingly, I enabled debug logging by setting datadog.logLevel: 'DEBUG" and I observer that the DD_LOG_LEVEL environment variable was correctly updated on the 'agent' and 'process-agent' containers, however not the 'trace-agent' container.

Describe what you expected:

• 'trace-agent' container to be removed from the DaemonSet

Steps to reproduce the issue:

1. Enable APM by setting datadog.apm.socketEnabled: true & datadog.apm.portEnabled: true.
2. Attempt to disable APM by setting datadog.apm.socketEnabled: false&datadog.apm.portEnabled: false`.

Additional environment details (Operating System, Cloud provider, etc):

• Using 3.7.6 version of Helm Chart
• Running in AWS EKS
• Using FluxCD for Helm deployments