Features

Monitoring

Seven ways to watch anything, from a public API to a cron job in your basement.

Every check type

HTTP(S), keyword matching, TCP ports, ping, DNS records, SSL certificates, and push heartbeats — one dashboard for all of it.

SSL checks that actually verify

Hostname, SAN, and full chain validation — not just an expiry date. A certificate about to expire shows as degraded before it becomes an outage.

Heartbeat (cron) monitoring

Give any cron job, backup script, or Uptime Kuma instance a private ping URL. When the pings stop, you get alerted — the failure your own stack can't self-report.

No false alarms

Confirmation thresholds require consecutive failures before a monitor is marked down, so a single network blip never pages you.

A degraded state, not just up/down

Warnings like a near-expiry certificate surface as degraded — visible and alerted, but never counted against your uptime.

Fast checks

Down to 1-minute intervals on Pro, 5 minutes on Free. Checks run on a fixed minute grid, so a 1-minute monitor really is checked every minute.

Automatic incidents

Incidents open and resolve themselves, with a full timeline, the detected cause, and the outage duration.

Response-time history

Per-monitor latency charts and hourly rollups, with uptime history up to a full year on Pro.

Alerting

Six channels, every one available on every plan — including free.

Alerts where you live

Email, ntfy, Telegram, Slack, Discord, and generic webhooks, the moment an incident opens or resolves.

First-class ntfy support New

Point alerts at ntfy.sh or your own self-hosted server. Down alerts arrive as urgent priority, recoveries as normal, and locked-down servers or reserved topics work with an access token.

Exactly-once delivery

A delivery ledger guarantees each incident transition reaches each channel exactly once — a flapping monitor can't double-page you or silently drop the recovery.

Flap suppression

A monitor bouncing up and down repeatedly gets its noise muted into paired notifications instead of a 3 a.m. barrage.

Channel health

Failing channels are tracked, retried with backoff, badged in the dashboard, and auto-disabled after repeated failures instead of silently eating alerts. Every channel has a one-click test send.

Recovery context

Back-up alerts include how long the outage lasted, so you know the blast radius before you open the dashboard.

Status pages

Tell your users before they tell you.

Branded public status pages

A clean, always-current page per product or team, with live status and the incident timeline.

90-day history bars

The classic per-day uptime strip, computed from real check data.

Custom domains

Serve your status page from status.yourdomain.com on Pro — with unlimited pages.

Public means chosen

Only monitors you explicitly add are shown, and a page is only public when you make it public.

Built for AI agents

Your agents can run your monitoring, not just read about it.

First-party MCP server

Claude and other MCP clients can create monitors, check current status, and act on incidents with a scoped API token.

Security & account

The boring parts, done properly.

Two-factor authentication

Authenticator-app 2FA with recovery codes, plus verified email changes on every account.

Encrypted at rest

Notification credentials, webhook URLs, and sensitive monitor settings are encrypted in the database, not just access-controlled.

Guarded outbound traffic

Every user-supplied URL we call — probes and webhooks alike — is SSRF-guarded, and alert content never leaks credentials embedded in monitor URLs.

Team roles

Destructive actions and billing are owner-only; teammates can operate monitors without being able to delete the account out from under you.

Teams & billing

One flat price per team. No asterisks.

Flat per-team pricing

No per-seat, per-monitor, or per-status-page charges. Free covers 10 monitors forever; Pro is one flat price for the whole team.

Self-serve billing

Upgrade, downgrade, invoices, and card changes through the Stripe billing portal — no email to support required.