Hi Product Hunt 👋 I built Trinetra because monitoring a handful of Linux servers kept turning into running a monitoring stack: a time-series database, an agent, a dashboard and an alert router, each with its own config and upgrades. Trinetra is one small daemon per server. It collects, stores 30 days of history, decides when something is wrong and tells you, on its own. Measured on a real box it uses about 11.6 MB of memory and ~0.4% of one core. The core is Go standard library only, enforced by a test. What you get: • Live dashboards for CPU, memory, disks, containers, systemd services and processes • Alerts to Telegram, Slack, Discord, email, ntfy, Gotify or a webhook, with "Who gets what" routing, quiet hours and silences • An optional web UI with passkey sign-in (no passwords) and roles • A public status page with named services and incident updates • Fleet mode: sudo trinetra fleet init on one server, a join token on the others; nodes stream to the master over mutual TLS and catch up after a network split • Updates signed twice (build pipeline and an offline maintainer key) that roll themselves back if the new build isn't healthy within 90 seconds It's self-hosted and the source is on GitHub under FSL-1.1-ALv2. Free to run on your own servers. I'd love to hear what you monitor today and what would make you switch. Happy to answer anything in the comments.

Trinetra - Self-hosted server monitoring in one ~12 MB daemon

Replies