Describe the enhancement:
When a user deploys Elastic Agent and configures Fleet enrollment using environment vars (common w/ Kubernetes), the Elastic Agent pulls the environment variables into local state and from that point on ignores environment variables.
This leads to some confusing behavior where a customer makes a mistake in providing the fleet URL or enrollment token, deploys the agent, sees the issue, fixes the issue, but the deployment continues to use the old values.
It would be nice if Elastic Agent would log that the provided environment variables do not match the current configuration and that the environment variables are ignored. Ideally we'd mention how to clear the current configuration by wiping the state.
Related to #3586 and https://discuss.elastic.co/t/agent-fleet-url-not-taking-what-is-in-the-yaml-config/370278/13