Project · 1 min read

Magpie

July 17, 2026

Once a collector fleet outgrows manual configuration management, a bad config push can silently take down telemetry for everything behind it — and nobody notices until the dashboards go quiet.

Why this exists

Application deploys earned guardrails years ago — staged rollouts, health checks, automatic rollback. Collector configuration mostly didn’t: on many fleets it still ships as “push and hope.” Magpie’s premise is that collector config deserves application-grade guardrails, because when telemetry breaks, it takes your ability to see everything else down with it.

What it does

• Fleet visibility — a live view of every agent in the fleet: which config version each collector runs, and the state of its pipelines.
• Safe remote configuration over OpAMP — config push and version tracking through the OpAMP protocol rather than out-of-band file drops.
• Canary rollouts with automatic rollback — changes go to a canary cohort first; a health regression reverts the change without operator intervention.

Status

Apache 2.0, solo-maintained, under active development. Python and FastAPI, containerised, built for the point where fleet size outgrows manual configuration management.