Uptimer v2.0.0-preview

Self-hosted monitoring that explains every decision: Resources, Signals, Rules and Incidents.

These pages describe the Uptimer 2.0 release candidate for the field test. Commands use the preview image ghcr.io/myuptime-info/uptimer:2.0.0-rc1. Uptimer 1.x stays the current release: its documentation is at /latest/.

Uptimer watches your services and tells you, with its reasons, when one is in trouble.

A Resource is something you watch: a website, an API, a job.
A Signal is one stream of evidence about it. A managed worker or your own scripts send Observations on it.
A Rule reads the Signals and decides whether the Resource is in trouble.
An Incident is one period of trouble from one Rule. It has a lifecycle (open, closed), a confirmation (confirmed, unconfirmed) and a history.
A Workspace holds Resources and their people. Everyone gets one, named Default, at first sign-in; rename it and invite members under Settings.

The field-test path

Quickstart: run Uptimer and one worker with Docker Compose and watch a website until it opens an Incident.
From an API key to an Incident: create a Resource, send an Observation and read the Rule result and the Incident with API v3 and the local Python SDK wheel.
Run it for real: PostgreSQL, OIDC sign-in, workers and their certificates, logs and Sentry.
Coming from 1.8: 2.0 is a fresh installation.

AI tools can use Uptimer through the installed binary’s MCP mode, and search these pages through the docs MCP.

Getting startedRun Uptimer 2.0.0-preview and complete one Resource-to-Incident path. OperatingOperate the Uptimer 2.0 control plane and its workers. ReferenceAPI v3, the Python SDK 2.0, the installed-binary MCP, and the docs MCP.