Launching today
Epure
Open source error tracker that runs on your VPS.
2 followers
Open source error tracker that runs on your VPS.
2 followers
Exception-only error tracking you run yourself. Keep official Sentry SDKs; point the DSN at your box. One Rust binary plus PostgreSQL 16. Two containers, no Redis, no Kafka. Grouped issues, breadcrumbs, releases, alerts, webhooks. Keyboard triage (j/k, e resolve, i ignore). Spike valve under loops: the counter still increments, duplicate bodies drop. No APM, replay, profiling, or log ingest. JS/TS sourcemaps. Other languages group on SDK frames. Apache 2.0. docker compose up.
Hey Product Hunt !
I'm Joshua, maker of Epure.
I kept wanting the same thing on small projects: when production throws, I want a grouped issue and a stack I can read. I already had official Sentry SDKs in the apps. What I did not want was standing up self-hosted Sentry's multi-service fleet just for exceptions, or waking up to an overage bill when one bug loops.
So I built Epure. Apache 2.0. Self-hosted. Exception-only.
Who it's for
• Indie SaaS, side projects, small startups on a VPS
• People who already use sentry/* and just want to change the DSN
• Anyone who wants calm triage without a full observability suite
What you get
• Keep your official Sentry SDKs, paste a new DSN
• Two containers: Rust app + PostgreSQL 16 (no Redis, no Kafka, no ClickHouse)
• Grouped issues, breadcrumbs, releases, alerts, keyboard triage
• Spike protection when one fingerprint floods
Out of scope on purpose
• No tracing, session replay, profiling, or log ingest
• JS/TS sourcemaps; other languages group on the frames the SDK sent
Try it:
Curious what you think: when you self-host error tracking, do you actually want APM and replay in the same box or is exceptions-only enough if the stack stays tiny?
Thanks for checking it out.
Happy to dig into any SDK or compose weirdness in the replies.