💡 WHY ALEX IS RIGHT
Paging on causes, like CPU, memory or a single failed health check, produces endless alerts that may or may not affect anyone. Paging on symptoms, measured as service level objectives on what users experience, asks one question: are we burning the error budget fast enough to break a promise? Fast burn pages a human; slow burn opens a ticket. One general rule replaces hundreds of hand-tuned thresholds and scales to every new service.
