create-heartbeat-monitor

Create a HEARTBEAT monitor

Preflight — read first. If you cannot see any uptimerobot:* MCP tools in your tool list, invoke the uptimerobot:setup skill before doing anything else. Do not tell the user the MCP is misconfigured — setup's Step 0 detects the common case (server connected, tools loaded after session start) and resolves it without re-keying.

Use when the user wants to monitor a scheduled job / cron / background worker / backup script. The target system must send periodic HTTP requests to a URL that UptimeRobot generates. If pings stop arriving for longer than the interval (plus optional grace period), the monitor goes down.

This is the inverse of the other monitor types: UptimeRobot does not probe anything, it waits for incoming pings.

Tool: create-monitor Required params: friendlyName, type: "HEARTBEAT". Do not send url. The server generates it.

Minimal call

{
  "friendlyName": "Nightly DB backup",
  "type": "HEARTBEAT",
  "interval": 86400,
  "gracePeriod": 1800
}

The response includes the generated URL, shaped like:

https://heartbeat.uptimerobot.com/m<id>-<hash>

Show that URL to the user so they can wire it into their cron / job.

Common optional params

• interval — how often the job is expected to ping (seconds, 30–86400). Required for this monitor to be useful; without it, defaults apply but may not match the user's schedule.
• gracePeriod — seconds of tolerance past interval before marking down (0–86400). Use this for jobs with variable duration.
• assignedAlertContacts, tagNames, maintenanceWindowsIds — same as other types.

Example: hourly job with 5-minute grace

{
  "friendlyName": "Hourly cache warmer",
  "type": "HEARTBEAT",
  "interval": 3600,
  "gracePeriod": 300,
  "assignedAlertContacts": [
    { "alertContactId": "10", "threshold": 0, "recurrence": 60 }
  ],
  "tagNames": ["jobs", "backend"]
}

Using the generated URL

The target job must hit the URL to signal "I ran successfully". Shell example:

curl -fsS --retry 3 "https://heartbeat.uptimerobot.com/m800123456-abcd1234" > /dev/null

Trigger on success only — if the job fails, skip the ping and UptimeRobot will raise an incident after interval + gracePeriod.

Common mistakes

• Sending url — HEARTBEAT rejects it (-29001). The server generates the URL.
• Sending timeout, port, httpMethodType, or any type-specific HTTP/KEYWORD fields — all rejected as forbidden fields.
• Setting interval shorter than the job's actual runtime, causing false downtime. Use gracePeriod to absorb normal variance.
• Forgetting to show the generated URL to the user. Without it they cannot wire up the job.

After creation

Call get-monitor-details to confirm. The details response exposes the heartbeat URL as well.

Related

• create-api-monitor — if the job also exposes a health endpoint you can probe.
• manage-monitors — for pausing during planned maintenance.