[AZ-302] C7 ThermalStatePublisher — jtop/NVML 1 Hz background poller

Implements AZ-297 InferenceRuntime's thermal_state() side: a singleton
background-thread publisher that polls jtop (preferred) or pynvml
(fallback) at config.thermal_poll_hz, stores an atomic ThermalState
snapshot, and emits c7.thermal_transition FDR records on every throttle
flip with a WARN log on entry and an INFO log on exit. Default-safe on
TelemetryUnavailableError per Invariant I-6 with a 1-Hz rate-limited
WARN.

Sources return a raw ThermalReading; the publisher stamps measured_at_ns
via its injected Clock so _JtopSource / _PynvmlSource stay clean of
direct time.* calls (Invariant 2). _poll_once is the deterministic test
seam — start() spawns the production thread.

- c7.thermal_transition registered in fdr_client.records KNOWN_PAYLOAD_KEYS
- [telemetry] optional dep group (jetson-stats, pynvml) added to pyproject
- 14 unit tests (AC-1..AC-6, AC-8, NFR-default-safe, structural)
  green; AC-7 / AC-1 microbench / NFR-perf-poll Tier-2 deferred
- full unit suite: 1140 passed, 11 expected Tier-2/CUDA skips

Co-authored-by: Cursor <cursoragent@cursor.com>

pyproject.toml

@@ -64,6 +64,14 @@ inference = [
     "onnxruntime>=1.17",
     # tensorrt is installed out-of-band on Jetson — not a pip dep
 ]
+# AZ-302: thermal telemetry backends used by C7's ThermalStatePublisher.
+# Both are Jetson / NVIDIA-host-only and not import-required for Tier-1;
+# the publisher selects whichever is importable at start() time and
+# raises TelemetryUnavailableError if neither is present.
+telemetry = [
+    "jetson-stats>=4.2",
+    "pynvml>=11.5",
+]
 indexing = [
     "faiss-cpu>=1.7",
 ]