When to use it
- Website or API availability and response time
- Health-check endpoints (
/health) - Web services that need custom headers, a request body or a specific status code
- Tracking SSL certificate and domain expiry for a site (see below)
Create an HTTP monitor
1
Enter the URL
Enter the Website URL, including
https:// or http://. The Monitor Name is optional and defaults to a name generated from the URL.2
Set how often to check
In How often we check, choose the Check Interval. The shortest interval depends on your plan; see Plans & Billing.
3
Choose probe locations
UptimeIO checks from multiple probe locations. On Pro and Scale you can choose which locations run your regular checks (at least one). On Free, locations are selected automatically.
4
Configure the request
Open HTTP Configuration to set the method, expected status, headers, body and timeout.
5
Attach notification channels
Choose where alerts go. See Notifications.
Settings
JSON checks are under Assert JSON response (see below). SSL certificate, domain expiry and slow response alerts are under SSL, domain & performance alerts (see below).
Example
- Website URL:
https://api.yourcompany.com/health - Method:
GET - Expected Status:
200 - Headers:
Accept: application/json - Check Interval: 5 minutes
Status codes and redirects
A check succeeds only when the response status is one of the Expected Status codes. By default only200 counts as success. Add other codes explicitly, for example 200, 201 and 204. Each code is listed individually.
Assert JSON response
Open Assert JSON response to check values inside a JSON response body. Each assertion has a path, an operator and, for most operators, a value. The check passes only when the response is valid JSON and every assertion passes. You can add up to 10 assertions, on every plan. A path uses dots for object keys and brackets for array positions, for examplestatus, data.items[0].id or [0].name. Keys start with a letter or underscore and may contain letters, digits, _ and -. Paths are up to 200 characters.
For equals, does not equal and contains,
true, false, null and numbers are read as that type. Wrap text in double quotes to force a string: "42" matches the string 42, while 42 matches the number.
Examples:
- Health endpoint returns
{"status": "ok"}: pathstatus, equals, valueok. - Queue size must stay low in
{"queue": {"depth": 12}}: pathqueue.depth, less than, value100. - Search returns at least one result in
{"items": [{"id": 1}]}: pathitems[0].id, exists.
Authentication
In HTTP Configuration, choose an Authentication type:- None (default)
- Basic Authentication: enter a username and password. UptimeIO sends them as an
Authorization: Basicheader. - Bearer Token: enter the token. UptimeIO sends it as
Authorization: Bearer YOUR_TOKEN.
X-API-Key: YOUR_SERVICE_KEY. Credentials are sent with every check, so use an HTTPS URL.
Slow response alerts
Under SSL, domain & performance alerts, turn on Slow Response Alert and enter a threshold in milliseconds (100 to 30,000). UptimeIO opens a separate slow response incident when a check takes longer than the threshold. It resolves when the response time stays below 80% of the threshold for 3 consecutive checks. A monitor can be up and still have an open slow-response incident.SSL certificate monitoring
Turn on SSL Certificate Monitoring for HTTPS URLs to track certificate health.
What happens:
- Warnings: at each selected threshold a warning is sent to the monitor’s notification channels. Warnings do not open an incident, because the site is still up.
- Expired or invalid certificate: opens an incident, which resolves automatically once a valid certificate is served.
Domain expiry monitoring
Turn on Domain Expiry Monitoring to be warned before the domain registration runs out. Domain expiry monitoring is available on Pro and Scale.- Available for HTTP, Keyword and DNS monitors.
- The URL must be on a publicly registered domain. IP addresses,
localhostand internal names are not accepted. - Choose the thresholds (30, 15, 7 and 1 days before expiry). All four are selected by default.
- Warnings go to the monitor’s notification channels. They do not open an incident.
- Subdomains are checked against their registrable domain:
app.example.comis checked asexample.com. A subdomain of a shared platform domain such asmyapp.vercel.appis checked against the platform’s domain, not yours. - Some country-code registries do not publish an expiry date. Those domains show as not published and never produce warnings.
Response time breakdown
Each check records DNS, TCP connect, TLS handshake, time to first byte (shown as Server) and total time. The monitor page shows the split per location, with the remainder as Transfer. Use it to tell whether slowness comes from DNS, the network, TLS or your application. See Reading metrics.Best practices
- Point the monitor at a lightweight health endpoint that checks your critical dependencies and answers quickly.
- Keep the timeout close to what a healthy response needs (5-10 seconds for APIs).
- Prefer
HEADwhen you only need availability and your server answers it correctly. - Expect incidents to open only after confirmation from several probe locations (see Understanding incidents).
- Allow the
UptimeIO-Monitor/1.0user agent through your firewall or WAF.
Troubleshooting
Timeout errors
Timeout errors
The server was too slow or unreachable. Check the timing breakdown to see which stage is slow, raise the Timeout (maximum 60 seconds, and always shorter than the check interval) if the server legitimately needs longer, and make sure a firewall or WAF is not blocking the
UptimeIO-Monitor user agent.Unexpected status code
Unexpected status code
Only
200 passes by default. Add the codes your endpoint returns to Expected Status. If the endpoint redirects, turn on Follow redirects or add the redirect code. Check with curl -I https://your-url.A JSON assertion fails
A JSON assertion fails
Check that the endpoint returns valid JSON, that the path matches the real response (open the URL in a browser or run
curl), and that the value has the right type. "1" (text) and 1 (number) are different values.SSL certificate errors
SSL certificate errors
An HTTPS check fails when the certificate is expired, self-signed, has an incomplete chain or does not match the hostname. Fix the certificate, and check that the chain includes intermediate certificates. For a test system with a self-signed certificate you can clear Verify SSL certificate under Advanced Options; do not do this in production.
Cannot choose probe locations or a short interval
Cannot choose probe locations or a short interval
Choosing locations and intervals under 5 minutes requires Pro or Scale. See Plans.
Next steps
Keyword Monitoring
Check page content as well
Notifications
Configure alerts for HTTP monitors
Related guides
Get started
Start monitoring your site free
Monitor your website or API, with SSL certificate expiry warnings included on every plan and domain expiry warnings on Pro and Scale. Create a free account to begin.