# Monitors

Source: https://getuptimely.com/docs/monitors

> Active probes, passive heartbeats, and how checks turn into incidents.

A monitor watches one thing and keeps a status timeline for it. Checks run
from Uptimely's probe runtime on a per-minute sweep; when a check result
crosses the criteria you configured, the monitor changes state and can
declare an incident, fire alerts, and notify subscribers — automatically.

## Monitor types [#monitor-types]

The create wizard offers, among others:

* **Website** — HTTP(S) checks with status-code, response-time and keyword
  criteria.
* **API** — like Website, plus request options: method, headers, body and
  redirect behavior.
* **Ping / IP** — reachability checks against a host. The check opens a TCP
  connection to port 80 and counts a completed connection as reachable; it
  sends no ICMP. For any other port, use a Port monitor.
* **Port** — TCP connect checks against a specific port.
* **DNS** — query one record type and fail when the query errors or comes
  back empty. It checks that the record resolves, not what it resolves to.
* **SSL Certificate** — certificate validity and expiry. See
  [what an SSL certificate monitor checks](https://getuptimely.com/features/ssl-certificate-monitoring).
* **Domain** — confirm the domain still exists, by DNS and then by WHOIS. It
  does not read the registration expiry date.
* **External Status Page** — watch a third-party provider's status page.
* **Synthetic / Custom JavaScript** — scripted checks.
* **SNMP** — poll SNMP-speaking devices.
* **Incoming Request** — a passive heartbeat: your job pings Uptimely, and
  silence beyond the expected interval is the failure signal.
* **Manual** — a status you set yourself, useful for composing status pages.
* **Data Check** — watch what an upAPI operation returns and get notified when it
  changes. See [Data Check monitors](/data-check).

## Request timeout [#request-timeout]

The probe request timeout is per-monitor configurable the way UptimeRobot
does it: **default 30 seconds, range 1–60 seconds**. It's editable on Website
and API monitors — in the create form ("Request Timeout (Seconds)") and later
on the monitor's Request Options tab. Other monitor types use the same
30-second default, non-configurable.

## Check interval and failure threshold [#check-interval-and-failure-threshold]

Each monitor runs on its own check interval, as often as every minute on
every plan. A failed check, once the re-check described under flap suppression
confirms it, changes the monitor's status straight away, so the status timeline
and status pages stay truthful. Declaring an incident or alert
waits for the monitor's failure threshold: **2 consecutive failed checks by
default, configurable from 1 to 10** on the monitor's Interval page. Recovery
has no threshold; the first passing check resolves what is open.

## Criteria and status [#criteria-and-status]

Each monitor carries criteria that map check results to a status (Operational,
Degraded, Offline) and decide whether an incident or alert is declared and
with which severity. Criteria are edited on the monitor's Criteria tab.

## Flap suppression [#flap-suppression]

A result that would move a monitor to a WORSE status is confirmed by an
immediate re-check before it commits — a single flaky read doesn't page
anyone. Recoveries commit immediately.

## Incoming-request heartbeats [#incoming-request-heartbeats]

Heartbeat monitors give you a unique URL; anything that can send an HTTP
request can report in. If the expected interval passes without a request, the
monitor degrades — the standard pattern for cron jobs and background workers.
