Self-hosted uptime monitoring
Know before
your users do.
OneUp watches your websites, APIs, servers, DNS and background jobs from your own infrastructure. When something slips, it opens an incident and tells the people who can fix it.
Acme production
All systems operational
Monitors
12
Uptime
99.98%
Response
118 ms
Response time
avg 118 ms
API · production
api.acme.com/health
Marketing site
acme.com
Checkout flow
acme.com/checkout
Postgres · primary
db-1.internal:5432
Nightly backup
heartbeat · 24h schedule
90-day uptime
99.98%
Recent activity
Incident resolved
Checkout flow · 6m ago
Latency recovered
API · 2h ago
Certificate valid
acme.com · 41 days left
7
probe types
10s
fastest check
5
alert channels
1
server to run
Monitoring
Seven ways to check that it is up
Every monitor has its own interval, timeout, retries and failure threshold. Checks land in the same timeline, so an API, a certificate and a nightly job read as one story.
HTTP / HTTPS
Request a URL and validate status codes, headers and response time.
Keyword
Assert that specific text is present, or absent, on a page.
Ping
Measure packet loss and round-trip latency to a host.
TCP port
Verify a port accepts connections and measure connect time.
DNS record
Resolve A, AAAA, CNAME, MX, TXT and NS records, then compare values.
SSL certificate
Track issuer, validity and expiry before a certificate lapses.
Domain expiry
Watch registration expiry through RDAP and renew on time.
Retries & thresholds
Require repeated failures before an incident, and one recovery to close it.
Monitors can be paused, duplicated, grouped into status-page components and exported as CSV on paid plans.
Create a monitorGetting started
From an empty server to a live status page
No setup wizard and no external account. Create the instance, add a check, and publish the page your customers will bookmark.
Create your account
The first account becomes the instance administrator. A workspace is created with it, ready for monitors and pages.
Add a monitor
Choose HTTP, keyword, ping, TCP, DNS, SSL or domain expiry. Set the interval, retries and who gets notified.
Publish a status page
Group components, pick a theme, connect a custom domain and share one URL customers can trust.
Incidents and alerts
One incident, not twenty alerts
A failing check does not page someone on the first miss. OneUp retries, waits for your threshold, then opens a single incident and keeps every update in one place.
- Configurable failure and recovery thresholds per monitor
- Incident updates with investigating, identified, monitoring and resolved states
- ntfy, email, Slack, Discord and generic webhooks
- Maintenance windows that suppress alerts and show on the status page
- Automatic recovery notices when the service comes back
Checkout latency above 2s
INC-204 · checkout flow · resolved in 13m
Investigating
14:02Checkout latency above 2 seconds on the p95.
Identified
14:06Database connection pool saturated during the deploy.
Monitoring
14:11Pool increased. Latency back under 300 ms.
Resolved
14:15Incident closed after 13 minutes. Status page updated.
Public status pages
A status page you would actually publish
Give customers one calm place to check. Components, 90-day uptime, incident history, maintenance windows and email subscriptions, all served from the same workspace.
Acme status
status.acme.com
All systems operational
Updated 8 seconds ago
API
Request routing and authentication
Web app
Dashboard and status pages
Checkout
Payments and order confirmation
Database
Primary Postgres cluster
Subscribe to updates
Self-hosted
Your monitoring data stays on your server
OneUp ships as prebuilt multi-arch images and runs with Docker Compose. Bring a server, a domain and a reverse proxy. There is no OneUp cloud account, no external analytics and no per-check bill.
One compose file
Postgres, Redis, API, workers and the web app.
Private by default
Database and Redis are not published to the host.
Instance admin
Manage users, workspaces, plans and suspension.
Your backups
A normal Postgres volume you can snapshot and move.
docker compose pull
docker compose up -d
# migrations run on bootDeployment
docker compose · oneup
Edge
Caddy · TLS · custom domains
Web
Next.js app · dashboard and status pages
API
FastAPI · auth, monitors, incidents, billing
Workers
Scheduler · monitoring · notifications
Data
PostgreSQL · Redis Streams
Plans
Start free. Scale when you need to.
Every plan includes monitoring, incidents, notifications, heartbeats, maintenance windows and public status pages. Paid plans add capacity and the features teams ask for next.
Plans are per workspace. On this self-hosted build an instance administrator assigns paid plans, and no card is charged by OneUp.
Starter
For side projects and the first things you want to keep online.
3 monitors with 60-second checks
1 public status page
1 heartbeat and 1 integration
1 workspace member
Incidents, notifications and analytics
Pro
For production services that need real alerting and a polished status page.
25 monitors with 30-second checks
5 status pages with custom domains
20 heartbeats and 10 integrations
5 workspace members
CSV exports and 50 components per page
Scale
For teams running many services, customers and scheduled jobs.
100 monitors with 10-second checks
15 status pages and white-label branding
100 heartbeats and 50 integrations
25 workspace members
200 components per status page
The instance administrator is not limited by a workspace plan. Limits are enforced by the API, and downgrading never deletes existing monitors or pages.
Questions
What to know before you start
The short answers. The README covers deployment, environment variables and the API in more detail.
Is OneUp really self-hosted?+
Yes. OneUp runs on your own server with Docker Compose. Postgres, Redis, the API, the scheduler, the monitoring worker, the notification worker and the web app all run in your environment. Nothing is sent to an external analytics or monitoring service.
How do checks and incidents work?+
A scheduler queues due checks, a monitoring worker runs the probe, and the result is evaluated against your retry and failure thresholds. Repeated failures open one incident instead of an alert per check, and a recovery closes it. Maintenance windows suppress alerts while work is in progress.
Can I use my own domain for a status page?+
Yes. Custom domains are available on Pro and Scale. Point the hostname at your Caddy or other reverse proxy, add the domain in the status page settings, and OneUp serves the page by hostname.
How are paid plans assigned?+
Plans are per workspace. This self-hosted build does not process card payments: an instance administrator assigns Starter, Pro or Scale from the admin console. Prices are shown for reference, and every limit is enforced by the API.
Start with one monitor.
You will know within a minute.
Create the instance, add your first check, and point a status page at it. The first account on the server becomes the administrator.