Define SLOs and runbooks for digest failures

**Phase:** 3 — Growth & Scale (Reliability)

- [ ] Define uptime/error budgets (e.g., 99% of daily digests sent within 2 hours of schedule)
- [ ] Write incident response runbooks for common failure modes:
  - Gemini API outage
  - SerpAPI quota exhausted
  - Supabase connection failures
  - Email delivery failures (Resend)
- [ ] Document escalation paths and recovery procedures