Pitched as Pingdom, PagerDuty and Datadog in one
OneUptime is a monitoring platform that runs uptime checks, incident management, on-call scheduling with escalation, status pages, and OpenTelemetry logs, metrics and traces from a single instance. The project pitches itself as the replacement for Pingdom, StatusPage.io, PagerDuty, Datadog and Sentry. It is written in TypeScript, installs with Docker Compose or Helm, and the company behind it also sells a hosted version, OneUptime Cloud.
The comparison of self-hosted uptime monitors and status pages weighs what OneUptime asks of a server against lighter tools.
SSO in the free edition, plus 30+ monitor types
- Detection to announcement in one place: the probe that catches the outage opens the incident, escalation pages whoever is on call by SMS, phone call, push, email or Slack, and the status page notifies its subscribers.
- More than thirty monitor types: HTTP, ping, port, SSL certificate, DNS, domain expiry, SQL query, scripted browser journeys in Chromium or Firefox, cron job heartbeats, SNMP devices, and even a vendor's status page such as AWS or GitHub.
- A status page wired to the probes: each row on the page is a monitor whose state updates on its own, while Cachet, in its released version, waits for statuses pushed through its API or entered in its dashboard. Subscribers hear about incidents by email, SMS, Slack, Teams or webhook.
- SSO in the free edition: SAML and OIDC ship with the self-hosted Community Edition. On the cloud they require the Scale plan, at $99 per user per month.
Expect seven containers and an upgrade every week
Three databases and seven containers. The Docker Compose install starts PostgreSQL for configuration and incidents, ClickHouse for telemetry and Valkey for cache and queues, plus the app, a probe, a runner and nginx. The docs recommend 16 GB of RAM, 8 cores and 400 GB of disk, with 8 GB and 4 cores as the floor for personal use. OneUptime does not handle TLS certificates, so a reverse proxy has to sit in front. For uptime and server health alone, Checkmate runs as two services, the app and MongoDB.
Kubernetes for production. The Docker Compose page advises against that setup in production and points to the Helm chart. The chart ships with no CPU or memory limits and 25 Gi volumes meant for a trial. For a small deployment, the sizing guide budgets 4 vCPU, 16 GB of RAM and 200 GB of NVMe for ClickHouse alone, which holds about 95% of the stored data.
Part of the code under a paid license since 14.0.0. Up to 13.0.8, the whole repository was Apache 2.0. Since 14.0.0, released on September 21, 2026, the ee/ directory falls under the OneUptime Enterprise License: SCIM, audit logs, team compliance, per-signal telemetry retention and the instance health dashboards need a per-user license in production once a 14-day trial ends. The Community image ships none of that code.
What the cloud adds, and what it bills. OneUptime Cloud runs the Enterprise Edition and probes from several regions. A self-hosted instance probes from its own server, so every extra vantage point is a probe you deploy yourself. SMS and voice alerts go through your own Twilio account, and mobile push goes through a relay on oneuptime.com by default. On the cloud, the free plan stops at one status page and 100 subscribers and bills each active monitor from $1 a month; on-call rotations need the Growth plan, at $22 per user per month.
Four major versions in just over three months. 11.0.0 came out on June 12, 2026, 14.0.0 on September 21 and 14.0.14 on October 4. The docs recommend updating at least weekly in production and say not to skip a major version, so an instance left alone for three months has to step through each one in turn.