Monitors
Check websites, APIs, TCP ports, Ping, SSL certificates and DNS records from the master.
A monitor is a check that the master runs by itself against an address: a website, an API, a database port, a certificate or a DNS record. It needs no agent, so you can use it for services on any server, including ones you don't manage.
Add a monitor
Open Monitors (the page is called Website / API monitoring) and click ＋ Add monitor.
| Field | What to enter |
|---|---|
| Name | For example "Activation service /health" |
| Group | The group it belongs to. Group alert recipients also get its alerts. |
| Runs on server | Optional: the server this service runs on. Hovering the monitor then highlights that server on the map. |
| Type | HTTP(S), TCP port, Ping, SSL certificate or DNS (see below) |
| Address | Depends on the type (see below) |
Before saving, click Test to run the check once and see the result. Save starts checking right away.
Types
| Type | Address | Up when |
|---|---|---|
| HTTP(S) | A URL, e.g. https://example.com/health | The response has the expected status code and, optionally, contains a keyword or a JSON value |
| TCP port | host:port, e.g. db.example.com:3306 | A TCP connection can be opened |
| Ping | A host name or IP, e.g. example.com | The host answers an ICMP ping |
| SSL certificate | example.com or example.com:8443 (port 443 if omitted) | The certificate is valid and has enough days left |
| DNS | A domain name, e.g. example.com | The name resolves, optionally to an expected value |
HTTP options
| Option | Default | Notes |
|---|---|---|
| Request method | GET | GET, HEAD, POST or PUT |
| Expected status code | 200-399 | A code (200), a list (200,301), a range (200-299) or a class (2xx) |
| Slow response threshold (ms) | No limit | Slower responses are shown as a warning (yellow), not as down |
| Page must contain | — | Text that must appear in the response (case-sensitive; the first 2 MB are read). Tick Invert: must not contain to alert when the text does appear, for example an error message. |
| JSON field and Expected JSON value | — | For APIs: a path such as status, data.ok or items.0.state (numbers are array positions). If you give a value, it must match (not case-sensitive); otherwise the field only has to exist. |
| Do not follow redirects | Off | By default up to 10 redirects are followed |
| Do not verify the certificate | Off | For self-signed certificates |
| Request headers | — | One per line, for example Authorization: Bearer xxx |
| Request body | — | Sent with POST and PUT |
Requests are sent with the User-Agent 3RMonitor/VERSION (uptime check). If a firewall or WAF in front of the site blocks unknown clients, allow it. A 403 caused by a Cloudflare challenge is reported as such.
SSL options
| Option | Default |
|---|---|
| Alert below (days left) | 10 days |
| Warn below (days left) | 30 days |
| Do not verify the certificate chain | Off; turn on for self-signed certificates |
DNS options
Record type: A / AAAA, CNAME, MX, TXT or NS. Expected to contain: optional text the answer must contain, for example 203.0.113.10.
Ping uses the system ping command on the master. If the master's network blocks ICMP, use a TCP port check instead.
Checks and alerts
| Field | Default | Meaning |
|---|---|---|
| Check interval (s) | 60 | How often to check (minimum 10 seconds) |
| Timeout (s) | 10 | Maximum 120 seconds |
| Retry interval after a failure (s) | 20 | Check again sooner while it fails: to confirm an outage quickly, and to notice the recovery quickly |
| Failures in a row before down | 3 | Several checks avoid false alarms (1 to 50) |
| Order | 0 | Position in lists |
| Enabled | On | Off = paused |
With the defaults, a site that goes down is confirmed about 40 seconds after the first failure (60-second interval, then two retries 20 seconds apart). During the outage it keeps checking every 20 seconds, so the recovery is noticed quickly too. You get one email when it goes down and one when it is back up; see Alerts and email.
Managing monitors
Monitors lists the monitors by group, with their type, address, status, the error of those that fail, and the last check. The tabs show All, Problems (down or with warnings) or Paused monitors, and Search name or address filters the list. Each row has these buttons:
- Check now: run the check immediately and show the result;
- Edit;
- ⋯ → Pause / Enable: a paused monitor is not checked and raises no alerts;
- ⋯ → Details;
- ⋯ → Delete: deletes the monitor and all its check results.
Click a monitor's name to see its details: availability over 24 hours, 7 and 30 days, the average response time, the days left on a certificate, the response time over 24 hours, 7 or 30 days (red areas are failed checks), the recent checks and its incident history.