create-http-monitor

Create an HTTP 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 website or HTTP(S) endpoint for reachability and/or specific response codes. HTTP is the default choice unless the user needs content matching (KEYWORD), response assertions (API), or low-level connectivity (PING / PORT).

Tool: create-monitor Required params: friendlyName, type: "HTTP", url (must be a full URL with scheme).

Minimal call

{
  "friendlyName": "Production API",
  "type": "HTTP",
  "url": "https://api.example.com/health"
}

Common optional params

• interval — seconds between checks (30–86400). Defaults to the plan's default. Free-plan minimum is higher than paid plans.
• timeout — request timeout in seconds (1–60).
• httpMethodType — one of HEAD, GET, POST, PUT, PATCH, DELETE, OPTIONS.
• authType — NONE (default), HTTP_BASIC, DIGEST, BEARER. Pair with httpUsername / httpPassword.
• customHttpHeaders — object, e.g. {"X-Source": "uptimerobot"}.
• successHttpResponseCodes — array of strings that count as success, e.g. ["200", "201", "2xx"].
• followRedirections — boolean, defaults to true.
• checkSSLErrors — boolean, validates cert errors when true.
• sslExpirationReminder / domainExpirationReminder — boolean flags for expiry alerts.
• responseTimeThreshold — milliseconds; 0 disables slow-response alerts.
• assignedAlertContacts — see below.
• tagNames — array of tag strings (max 20, ≤50 chars each).
• maintenanceWindowsIds — array of maintenance window IDs.

Alert contacts

"assignedAlertContacts": [
  { "alertContactId": "10", "threshold": 0, "recurrence": 0 },
  { "alertContactId": "20", "threshold": 5, "recurrence": 30 }
]

alertContactId values come from list-integrations. threshold = minutes of downtime before this contact fires. recurrence = re-notify every N minutes while still down.

Full example with auth + headers

{
  "friendlyName": "Internal API health",
  "type": "HTTP",
  "url": "https://internal.example.com/api/v1/health",
  "interval": 60,
  "timeout": 10,
  "httpMethodType": "GET",
  "authType": "BEARER",
  "httpPassword": "sk_live_...",
  "customHttpHeaders": { "X-Env": "prod" },
  "successHttpResponseCodes": ["200", "204"],
  "sslExpirationReminder": true,
  "assignedAlertContacts": [
    { "alertContactId": "10", "threshold": 0, "recurrence": 0 }
  ],
  "tagNames": ["prod", "critical"]
}

Common mistakes

• Passing a bare hostname ("api.example.com") — HTTP monitors require a full URL with scheme. Use PING for hostname-only checks.
• Sending numeric type: 1 — MCP only accepts the string "HTTP".
• Listing forbidden fields like port or gracePeriod on an HTTP monitor. The server rejects with -29001 invalid_parameters.
• Setting interval below the plan minimum. Returns -28002 subscription_limit_exceeded.

After creation

Call get-monitor-details with the returned monitor id to confirm stored config. Read-after-write can lag a few seconds.

Related

• create-keyword-monitor — if you need to assert specific text is (or isn't) in the body.
• create-api-monitor — if you need JSONPath assertions on the response body.
• errors — for recovering from -28002, -29001, 429.