Open Source
Self-hosted uptime monitoring.
Run it on your own boxes.
The open-source engine behind ObserveOne. Every monitor type, run from your own boxes. Multi-region. Apache 2.0.
What it does
Production monitoring you run yourself.
Same monitor types the SaaS uses, minus the SaaS. One image, one command.
Coverage
Every monitor type
HTTP and API checks with assertions. TCP and UDP reachability. Database liveness. TLS certificate health. Browser checks. Heartbeats. One worker runs them all.
Inverted ping
Heartbeats
For cron jobs and background workers. Your service pings a URL we give you on every successful run. We alert when a ping doesn't arrive in time. Recovery alerts fire when it comes back.
Multi-region
DistributedAgents pull work from the master. No inbound ports needed on the agent side. Spin one up wherever you need a check to run from.
View docs →
Dashboard included
UI includedSign in, add monitors, watch runs, set up status pages, wire alert channels. Built into the same image, same port. No extra service to run.
Alerts
AlertingWebhook, Discord, Slack, and email. Bind any channel to any monitor. Bring your own SMTP. No third-party alert vendor sits between you and the outage.
View docs →$## Investigating
$DB connection saturation in eu-west-1.
$
$## Identified
$Rolling back the 14:02 deploy.
$
$## Resolved
$Latency back to baseline.
Status pages with incidents
Public-facingPublic status pages list your monitors plus an incident timeline you write yourself. Markdown-rendered, XSS-safe, no third party. Backups carry the timeline along with everything else.
$curl -O https://monitor/runs/482/trace.zip
$npx playwright show-trace trace.zip
Failed runs come with traces
DebuggableEvery failed browser run saves the Playwright trace and a screenshot to bundled S3-compatible storage (RustFS, no setup). Download from the run detail page and open locally to see exactly where it broke. Backups bring them along to the new host.
$bun scripts/export.ts --include-artifacts -o backup.oodump.tar.gz
$bun scripts/import.ts --from backup.oodump.tar.gz --force
Backup & restore
PortableOne command dumps the database and every Playwright artifact into a single tar.gz; one restores it on a fresh host. The schema-version guard refuses dumps that don't match the target's migration head, so you can't silently downgrade.
View docs →${
$ "warnDays": 30,
$ "verifyChain": true,
$ "verifyHostname": true,
$ "expectCnRegex": "^api\\.prod\\."
$}
Deep TLS checks
TLSBeyond expiry: assert the full chain trusts a public root, the hostname matches the certificate's SAN, and the CN/SAN matches a regex you control. A reverse-proxy swap that quietly serves the wrong cert fails the check instead of slipping into prod.
View docs →