Add observability for lock poison-recovery and new bulk HNSW build paths

[coderabbitai]

Summary

Silent lock-recovery and the new bulk-build paths alter runtime behaviour without new logs, metrics, or spans at the decision points.

Required changes

• Emit a warning/metric or tracing span when poison recovery occurs (e.g. in `DistanceCache`, `RecordingLayer`, `FixedMonotonicClock`, `init_logging`).
• Add a build-level span/latency metric for the new bulk HNSW constructors (`CpuHnsw::build`, `build_with_edges`, `with_capacity`).

Affected areas

• `chutoro-core/src/hnsw/distance_cache.rs`
• `chutoro-core/src/hnsw/cpu/construction.rs`
• `chutoro-test-support/src/lib.rs`
• `chutoro-cli/src/logging.rs`
• `chutoro-core/src/session/clock.rs`

Acceptance criteria

• Poison-recovery events are observable (log/metric/span).
• Bulk HNSW build paths emit latency telemetry.

References

• PR: $pr_url
• Requested by: @$assignee

[coderabbitai]

🔗 Related PRs

#122 - CPU session scaffold: SessionConfig, ClusteringSession, build_session [closed]
leynos/axinite#195 - Use monotonic build duration clock seam (#194) [closed]
leynos/podbot#100 - Add cargo audit dependency gate [closed]
#130 - Refresh repository documentation guidance [merged]
leynos/lading#133 - Count publish index-lookup downgrade events (#68) [closed]

---

📝 Issue Planner

_{Check the box below or use the @coderabbitai plan command to generate an implementation plan and prompts that you can use with your favorite coding assistant.}

• Create Plan

---

🧪 Issue enrichment is currently in open beta.

You can configure auto-planning by selecting labels in the issue_enrichment configuration.

To disable automatic issue enrichment, add the following to your .coderabbit.yaml:

issue_enrichment:
  auto_enrich:
    enabled: false

💬 Have feedback or questions? Drop into our discord!