Who monitors the monitoring? Design meta-monitoring so a dead collector never hides an outage.

Monitoring dies silently: stuck alert queues, full TSDBs, collectors killed inside the same blast radius. Layer independent checks and a heartbeat so absence of signal pages louder than any red dashboard.

Grounded in researched DevOps, SRE and platform engineering interview loops, written to a senior-engineer editorial bar, and never padded to hit a word count.