Upright
A free, self-hosted tool that runs automated checks on your websites from multiple locations and tells you the moment something breaks — before your customers notice.
🔗 Visit Upright
Description
Most uptime monitors are a paid subscription to someone else's cloud, watching your site from servers you don't control. Upright flips that: it's a piece of software you run yourself, on cheap servers you pick, that pokes at your websites and services on a schedule and screams (via alerts) when something's wrong.
Upright is the open-source synthetic monitoring system that Basecamp built to watch over Basecamp, HEY, Fizzy, and their other products. It's a Rails engine you deploy to inexpensive VPS nodes around the world using Kamal, running health-check probes — real browser checks via Playwright (which can log in, click through a flow, and verify it actually works end-to-end, capturing video on failure), plain HTTP checks, SMTP handshake tests, and traceroute latency analysis. Everything reports back as Prometheus metrics and OpenTelemetry/OTLP data, so it plugs into AlertManager for paging and Grafana for dashboards, with OmniAuth login for any OIDC provider.
💬 Our review
The short version: if you're comfortable running your own infrastructure, Upright gets you real, multi-location, Playwright-powered uptime monitoring — including actual login-and-click-through checks, not just "did the homepage return a 200" — for the price of a couple of small VPS boxes instead of a monthly SaaS bill.
The catch is exactly what makes it free: you own the ops. There's no managed dashboard, no vendor support line, and no polish beyond what Basecamp needed for their own stack — you're deploying a Rails engine, wiring up Prometheus/AlertManager/Grafana yourself, and running nodes on providers like DigitalOcean or Hetzner (minimum 2 vCPU / 2GB RAM / 25GB disk each). For a team with the DevOps chops and appetite for self-hosting, that's a genuinely good deal against paying Datadog- or Checkly-style prices for synthetic checks. For anyone who just wants an uptime alert without touching a server, a managed tool like UptimeRobot or Checkly will get there in minutes instead of an afternoon of Kamal deploys.
💰 Pricing
📊 Global score
🤖 AI-enriched data
Licence MIT, auto-hébergé — le seul coût est l'infrastructure (VPS ~2 vCPU/2GB RAM/25GB disque minimum par nœud).
Pros
Gratuit et open-source (licence MIT), aucun abonnement récurrent
Checks navigateur réels via Playwright — login, parcours complet, capture vidéo à l'échec
Export Prometheus/OpenTelemetry natif, s'intègre à Grafana et AlertManager déjà en place
Déploiement multi-site géographique via Kamal sur n'importe quel VPS
Cons
Aucun tableau de bord managé ni support fournisseur — 100% à la charge de l'équipe
Nécessite de savoir déployer/opérer soi-même (Rails, Kamal, Prometheus)
Pas de version cloud clé-en-main pour qui veut juste une alerte en 5 minutes