Elastic Agent Stays in "Healthy" State when it has been unenrolled/API Key invalidated

[BenB196]

For confirmed bugs, please report:

• Version: 8.17.1
• Operating System: Container/Linux
• Discuss Forum URL: N/A
• Steps to Reproduce:
  1. Create an Elastic Agent fleet policy
  2. Enroll/install an Elastic Agent against the policy
  3. Wait for the agent to check-in and become "Healthy"
  4. Unenroll the agent
     • This can be done manually via Kibana UI/API
     • This can happen automatically if the Agent policy has an "Unenroll" time set
  5. Observe the following:
     • Elastic Agent logs "API key has been invalidated"
     • Elastic Agent when running elastic-agent status reports as healthy
  6. If you restart the agent/host after, the agent will then enter an unhealthy state

I would expect that once the agent starts to fail check-ins or sees it's API key has been invalidate, it should report an unhealthy state.

I've randomly run into this issue a few times (mainly from intermittent network connectivity issues that trigger the agent policy unenroll), but it becomes an issue when trying to detect when an agent is in an "unhealthy" state for things like; elastic/cloud-on-k8s#5376

[elasticmachine]

Pinging @elastic/elastic-agent-control-plane (Team:Elastic-Agent-Control-Plane)