Story 3.5 — Monitor health: degraded state, `/status`, and alerting

**Epic:** #736 — BN traffic-shaper monitor

Design: [v4 design](https://github.com/automa-saga/traffic-shaper/blob/main/docs/bn-qos-multiclass-priority-design-v4-nft-priority.md) §8.1, §10 risk 3.

Surface monitor health so a crashed/stuck monitor (which silently loses ingress shaping on the next pod restart) is detected.

**Acceptance criteria:**
- Surface monitor state (including `degraded`) via `GET /block_node/traffic_shaper/status`.
- Alert on the monitor's `degraded` state, not only on pod events: a crashed monitor leaves stale qdiscs until the next pod restart, at which point ingress shaping is lost indefinitely.
- Document the daemonkit back-off behaviour and the operator response.

Start with docs/bn-qos-multiclass-priority-design-v4-nft-priority.md, especially sections 8.1 and 10 risk 3, then locate the monitor and its HTTP status entry point. Verify that GET /block_node/traffic_shaper/status exposes degraded state, that degraded state triggers an alert, and that the daemonkit back-off behavior and operator response are documented.

Written by the indexing model from the issue text.