Hi everyone. I've been on both sides of the Sentry tradeoff. On one project the cloud bill kept climbing every quarter. On another we self-hosted Sentry and spent more time babysitting Kafka, ClickHouse, Redis, and Symbolicator than reading stack traces. Both options felt wrong for small teams. So I built urgentry. The goal: keep every Sentry SDK working, ship the whole thing as one binary, and make evaluation cheap. If you can scp a file to a $5 VM, you can run it. Change one DSN, send one event, see if the switch works. If it doesn't, your old Sentry is still sitting there. The hardest part wasn't ingest or the UI. It was deciding what to leave out. Tiny mode runs on SQLite, so every query had to earn its keep. The split-role build (Postgres + NATS + S3, Valkey for shared state) is in the repo for teams that outgrow Tiny mode, but Tiny is the public starting point on purpose. Most of the Sentry value people use is issue triage, alerts, releases, and a few dashboards. Once those work end-to-end on one binary, the rest of the platform stops feeling required. I'd love honest feedback, especially from anyone who's tried self-hosting Sentry and walked away. What made you walk away? What would it take to give it another shot?

urgentry - Drop-in Sentry replacement

Sentry, minus the bill and the YAML

Sentry, minus the bill and the YAML