Incidents I: Slowness & Data-Layer Pain

War-room playbooks for slow systems: databases, connection pools, N+1, caches, Redis outages and a blocked Node event loop — symptoms, first 5 minutes, fixes, and the interview answer.

About 110 minutes · 15 exam questions · 9 interview questions · 2 code labs

What you will learn

  1. Incident 0 — "The whole system is slow": the 5-minute triage protocol
  2. Incident 1 — Database slow / DB CPU pegged at 95%
  3. Incident 2 — DB connections exhausted ("too many clients already")
  4. Incident 3 — Application connection pool exhausted (DB looks fine)
  5. Incident 4 — Too many DB queries: the N+1 problem
  6. Incident 5 — Cache missing (hit ratio collapse) and cache stale
  7. Incident 6 — Cache stampede and the hot key
  8. Incident 7 — Redis is down
  9. Incident 8 — One API is extremely slow (everything else is fine)
  10. Incident 9 — Event loop blocked / thread pool exhausted (Node.js)

Next: Incidents II: Scale & Growth