CSP.LMC Health Supervisor
==========================

Role in the Architecture
------------------------

The CSP.LMC Health Supervisor coordinates the supervised evaluation of the
aggregated system health for a CSP.LMC device.

It applies debounce and maximum-latency rules to incoming subsystem
events, ensuring that health evaluation is:

- Stable
- Predictable
- Resistant to transient fluctuations
- Executed within bounded time

The supervisor does not contain business evaluation rules.
Instead, it orchestrates when evaluation should occur and
delegates the actual health computation to the health model.

Responsibilities
----------------

The supervisor is responsible for:

- Receiving subsystem events
- Storing updates in a thread-safe structure
- Applying debounce and max-latency timing control
- Triggering snapshot-based evaluation
- Ensuring evaluation is not repeated unnecessarily
- Buffering forwarded subsystem ``HealthInfo`` updates (latest per
  subsystem source).
- Coalescing forwarded ``HealthInfo`` with locally evaluated diagnostics
  into a single publication per debounced cycle (subject to change
  detection).

It acts as the coordination layer between asynchronous
subsystem inputs and deterministic health publication.

Interaction with Other Components
----------------------------------
The supervisor operates between:

- The subsystem event sources
- The ``HealthStateModel`` (health classification)
- The ``HealthInfoManager`` (diagnostic aggregation and publication)

It coordinates evaluation timing and ensures that health classification
and diagnostic aggregation remain independent from concurrency
and scheduling concerns.


For detailed API documentation, refer to
:doc:`/api/observation/supervisor/health_supervisor`.