← All work
BackworkLOGS & MONITORING FOR YOUR APPS
Developer & infrastructureLivebackwork.dev

Backwork

A self-hosted alternative to Datadog or Sentry — one server that collects your apps' logs, errors and performance and shows what's actually broken right now, without a bill that climbs with usage.

One screen
everything that's breaking, together
Honest
tells you when monitoring itself breaks
Runs itself
on your own box, no metered bill

The problem

Monitoring tools get expensive fast, and most dashboards happily show 'all green' even when the data behind them has quietly stopped arriving. Backwork runs on your own server, watches the machine it's on, and is honest about the gaps.

What it does

It sits alongside your apps and collects everything they give off — the logs, the resource usage, the errors — and shows what's going wrong right now: crashes, out-of-memory kills, restart loops, spikes in errors. When the data stops arriving it says 'no data', instead of a comforting green light that isn't true.

It tracks how fast your app is responding and how often it's failing, alerts you the moment something breaks (on Slack, Discord, email or text), and folds repeat errors into one so the real problem isn't buried. Getting an app reporting into it is a quick add, not a rebuild.

Key features

  • Logs, metrics and traces in one place

    When something breaks you look in one dashboard instead of stitching together three separate tools under pressure.

  • Honest 'no data' status

    You find out when monitoring itself stops working — the thing most tools hide behind a comforting green — instead of trusting a false all-clear.

  • Alerts where your team already is

    The moment something breaks, it reaches you on Slack, Discord, email or SMS, so nobody has to be watching a screen.

  • Repeat errors grouped together

    A thousand copies of the same error collapse into one, so the real signal isn't buried under noise.

  • Runs on your own box, no metered bill

    A small team gets proper monitoring without a bill that grows with usage — and none of the data ever leaves their own servers.

Where it stands

Watches the host it's deployed on rather than a cross-server fleet, and has no public users yet.

Under the hood — for the technically-minded

How it's built

Backwork runs as its own stack alongside the services it watches and pulls their logs, metrics and traces directly into a single dashboard. Its own instrumentation SDKs and a small CLI mean onboarding an app is a light dependency rather than a rebuild.

It's multi-tenant from the ground up — organisations, teams and projects with isolated access, verified by a real cross-tenant isolation test — with threshold-based alerting over the channels a team already uses.

The hard problems

  • Telling the truth when telemetry stops

    Most dashboards keep showing 'all green' after data stops arriving — the most dangerous possible lie for a monitoring tool. Backwork deliberately shows 'no data' instead of a false healthy state, so a silent collector failure looks like a problem, not calm.

  • Observability without per-gigabyte pricing

    SaaS observability gets expensive fast. Backwork gives a small team the same core capabilities — logs, metrics, traces, alerting and error grouping — self-hosted, without a per-gigabyte bill, built first for the team's own infrastructure.

  • Auditing its own security before shipping

    It carries a genuinely thorough internal self-audit — across architecture, performance, security and multi-tenancy — that found and fixed real issues before launch, rather than waiting for someone else to.

Built with

  • Remix
  • Postgres
  • Docker
  • Self-hosted

Building something in this space?

Work with us →