Server Health Monitoring System 10M instances · 75M points/s ingest · absence detection · breach → page in ≤ 60 s · alerting stays up when storage is down
arrows 0/25cards 15letters A–N
happy-path side-effect failure-comp new-infra
▶ BOTE — canonical numbers (every N- chip in the rail links here)
▶ Failure playbook (A–N)
▶ What breaks first · interview framing