Add internal Microsoft detector health dashboard and regression metrics

### Is there an existing issue for this?

- [x] I have searched the existing issues

### Is your feature request related to a problem? Please describe the problem.

Event-weighted telemetry showed approximately 86% alias coverage for 13.5, while only 41% of positively detected GA devices had aliases. Ad hoc event-level queries therefore hide device coverage, high-volume outliers, source churn, and platform-specific probe failures.

### Describe the solution you'd like

Create maintained Kusto queries or a dashboard covering device-weighted internal rate, alias coverage, detector outcome, per-probe outcome and latency, OS, CI, cache behavior, physical versus identity version, and source churn. Include canonical alias derivation for affected historical VS Code values.

### Additional context

Parent tracking issue: #19645

Acceptance criteria:

- Device-weighted and event-weighted metrics are shown separately.
- Strong probes producing zero results and sudden source shifts are visible.
- Detector failures and timeouts can be filtered by version and platform.
- 13.5 and later versions can be compared without ad hoc query rewriting.

Depends on the detector outcome schema tracked by the instrumentation child issue.