FlareWatch checks your sites, APIs and cron jobs every minute and tells you when they break. It runs on Cloudflare Workers in your own account, so there's no server to look after. Fork it, edit one config file, push.
What it watches
Status codes, a keyword that must (or must not) appear, a response header, a JSON value, a timeout, and your own headers and body. A failed check can be confirmed from a second location before it counts, or you can wait for a few failures in a row. A slow one shows as degraded.
Databases, mail servers, game servers. The Worker opens a connection to the host and port.
db.example.com:5432 open
The job pings FlareWatch when it finishes. If the ping is late, never arrives, or reports a failure, you hear about it.
Cron job monitoring$ ./backup.sh; curl -fsS \
"https://status.example.com/ping/nightly-backup/<token>/$?"Days before a certificate expires, the monitor shows as degraded and you get one warning. An expired or invalid certificate fails the check, and so does a certificate in its last 24 hours through flarewatch-proxy. Needs Globalping or a proxy, because a Worker can't see the certificate.
Look up an A, AAAA, CNAME, MX, TXT, NS or CAA record over DNS-over-HTTPS, and check that it holds the values you expect.
FlareWatch reads the domain's expiry date over RDAP, once a day by default. When the domain expires within 30 days, or within the number of days you set, the monitor shows as degraded and you get one warning.
Run a check from Globalping probes in another country, or through flarewatch-proxy, a small proxy you run inside your own network. Or reach a private network through Workers VPC and a Cloudflare Tunnel.
Checked, saved and alerted like the rest, but kept off the public page and API.
Alerts
Each webhook can take only the monitors you name, and a monitor can remind you every 30 check runs or more while it stays down. When one check run has many alerts, a webhook can get one message that lists them.
Sends to
Reverse proxy is down
Also down: App, Wiki, Git, Photos, Docs
Reverse proxy is up
dependsOn: ['proxy'] on each appStatus page
90 days of uptime and incidents per monitor, and a 12-hour response time chart.
Hide single monitors, or make the whole page private.
A password, GitHub, or any OpenID Connect provider. Give teammates read access, or let a client see only their own monitors.
Planned work shows on the page, and monitors inside a window send no down alerts.
Post a notice that shows under the status banner. Incidents, maintenance and notices also go out in an Atom feed.
Edit the colour variables in one CSS file to match your brand, in light and dark.
/api/badge.svg?id=demo_example
/embed/demo_github_statusGET /api/data{
"up": 6,
"down": 0,
"monitors": {
"demo_example": {
"up": true,
"status": "up",
"latency": 5,
"location": "SIN",
"message": "OK"
},
…
}
}Deploy
Leave GitHub Actions on. The fork is your deployment.
Your Cloudflare account ID and an API token. The other secrets are optional, like alert webhooks you keep out of git.
Monitors and alerts go in worker.ts, page settings in public.ts.
The workflow runs the checks and deploys both Workers. The first deploy takes about ten minutes.
export const workerConfig: WorkerConfig = {
monitors: [
{ id: 'api', name: 'API', method: 'GET', target: 'https://example.com/health' },
{ id: 'db', name: 'Database', method: 'TCP_PING', target: 'db.example.com:5432' },
{
id: 'backup',
name: 'Nightly backup',
method: 'HEARTBEAT',
periodSeconds: 86400,
graceSeconds: 3600,
},
],
};[{ "url": "https://hooks.slack.com/services/...", "template": "slack" }]$ git commit -am "Add my monitors" && git push
✓ Deployed to flarewatch.<you>.workers.dev
Cost
| Free tier limit | What uses it |
|---|---|
| 100,000 Worker requests/day | Page views, API calls and heartbeat pings. |
| 100,000 Durable Object requests/day | Check runs (1,440 a day), page views, API calls and pings. |
| 5 million rows read/day | Check runs (about 9,000 a day, plus 1,440 per monitor) and page views (two rows per monitor each, plus a few). |
| 100,000 rows written/day | Check runs (about 2,900 a day), incidents, pings and edits. |
| 1,000 KV writes/day | Sign-ins. |
| 5 cron triggers per account | FlareWatch uses one. |
Limits
For a personal or small-team page, yes. Everything fits in Cloudflare's free tier. The limit you're most likely to hit is 100,000 Worker requests a day from page traffic.
A Cloudflare account on the free plan and a GitHub account. A custom domain is optional; without one the page lives on a workers.dev address.
Read the changelog for breaking changes first. Then open your fork on GitHub and select Sync fork, then Update branch. If the deploy doesn't start, run CI and Deploy from the Actions tab.
In your own Cloudflare account. Check history sits in a Durable Object, sign-in sessions in KV. Nothing passes through flarewatch.app.
Yes. Mark single monitors private, or make the whole page private so visitors only see a sign-in page. Heartbeat pings keep working either way.
Yes, through flarewatch-proxy, a small service you run inside that network, or through Workers VPC and a Cloudflare Tunnel. Globalping probes cover checks from other countries.
Fork the repo, add two secrets, list your monitors, push.