#alerting

Short answer: treat each metrics poll as a bounded, idempotent read, and treat HTTP 429 as an...

Some months ago I wrote here about one bad instance hiding inside a fleet average. This is the...

Short answer: for Node.js failure alerting, use API polling over scheduled runs, logs, and errors for...

Alert on a media experiment only when an unresolved error group crosses a cohort-specific decision...

A panel gets added mid-incident, nobody owns it, and a year later it's flat or wrong. Here's how to tell a dead metric from a live one.

Alert scheduled media imports with two signals: a failure counter for bad runs and a heartbeat for...

A healthtech alert is useful only if it preserves enough evidence to reconstruct the customer...

A backend SaaS error alerting API for a media notification service has an awkward constraint: failed...

The least complex useful result is a scheduled Node.js cron worker polling a metrics API for...

The four PostgreSQL numbers that predict an outage, the SQL to read them, sensible thresholds, and how to alert on them without installing an exporter.

The disk alert on our primary database volume fired at 14:02. It was correctly written, correctly...

A Prometheus rule that only fires while healthy, an exporter scraped for nothing, a curl -sf that succeeds on an HTML page, and a pager that rang for an upload timeout. Four monitoring false negatives, with the mechanism behind each.