Workstream: Observability (#30, #44, #45, #68)

[ariaxhan]

Observability Workstream

All issues related to making KERNEL's internal behavior visible and measurable.

Issues

Issue	What	Priority	Status
#30	Full observability — telemetry events, agent tracking, hook instrumentation	P1	Open (epic)
#44	Instrument all hooks with execution timing	P2	Open (depends on #30)
#45	/kernel:metrics — monitoring dashboard with actionable insights	P2	Open (depends on #30)
#68	R-factor continuous quality scoring	P2	Open

Consolidated into this workstream

• feat(agentdb): add health command — subsystem status check #46 (health command) — folded into feat(agentdb): full observability — telemetry events, agent tracking, hook instrumentation, health dashboard #30
• feat(hooks): session-start telemetry summary — replace static with dynamic insights #48 (session-start telemetry) — folded into refactor(hooks): enrich session-start with dynamic telemetry + learning highlights #24

Execution order

1. feat(agentdb): full observability — telemetry events, agent tracking, hook instrumentation, health dashboard #30 first — the telemetry events table and agent tracking are prerequisites for everything else
2. feat(quality): R-factor continuous quality scoring — replace Big 5 pass/fail with crystallographic metrics #68 parallel — R-factor scoring is independent, can ship alongside feat(agentdb): full observability — telemetry events, agent tracking, hook instrumentation, health dashboard #30
3. feat(hooks): instrument all hooks with execution timing #44 after feat(agentdb): full observability — telemetry events, agent tracking, hook instrumentation, health dashboard #30 — hook instrumentation writes to the events table from feat(agentdb): full observability — telemetry events, agent tracking, hook instrumentation, health dashboard #30
4. feat(commands): add /kernel:metrics — monitoring dashboard with actionable insights #45 last — metrics dashboard consumes data from feat(agentdb): full observability — telemetry events, agent tracking, hook instrumentation, health dashboard #30 and feat(hooks): instrument all hooks with execution timing #44

For agents

#30 is the foundation. If you're working on #44 or #45 and #30 isn't done, stop — you have no data source.