OneUptime

The single-server Docker Compose install recommends 16 GB of memory, 8 cores, and 400 GB of disk, and the project notes that a Raspberry Pi works for a homelab. A production Kubernetes install uses the Helm chart and three datastores: PostgreSQL for monitors, incidents, and users, ClickHouse for logs, metrics, traces, and exceptions (about 95 percent of storage), and Valkey for cache, queues, and sessions. Object storage is only needed if you enable object-store backups.

The dashboard runs in the browser and installs as a desktop app from Edge, Chrome, or Safari on Windows, macOS, and Linux. OneUptime On-Call is a native app for iOS 16.4 or later and Android 7.0 or later, with an APK available from GitHub. Automation runs through the @oneuptime/cli npm package, which talks to the REST API with a project API key.

OneUptime publishes migration guides for Pingdom, UptimeRobot, Better Stack, Statuspage, PagerDuty, Opsgenie, Incident.io, Grafana On-Call, Splunk On-Call, Uptime Kuma, and StatusCake, and its Import from another tool page reads a source account with a read-only API token. The Pingdom importer shows a preview of what it found, lets you tick what to bring over, and remembers imported IDs so a re-run only creates new records.

Every dashboard action is available through a REST API authenticated with a project-scoped API key, and the oneuptime CLI wraps the same resources for terminal and CI use. A Terraform and OpenTofu provider manages monitors, status pages, on-call policies, teams, and labels as code, and an MCP server exposes project records as tools for AI clients.

Users can sign in with a local password or a passkey, and projects can add SAML 2.0 or OpenID Connect single sign-on, including a setting that requires SSO for everyone. SSO is part of the Community Edition on self-hosted installs, while SCIM provisioning is an Enterprise feature. Multi-factor authentication is handled by your identity provider when SSO is enabled.