Terminal
Terminal/Skill
Modify an existing UptimeRobot monitor — rename, change URL/interval, swap alert contacts, retag, tweak HTTP/keyword/API/heartbeat settings, and toggle SSL or domain reminders.
By UptimeRobot
Works with
npx skills add uptimerobot/ai --skill update-monitorPreflight — read first. If you cannot see any
uptimerobot:*MCP tools in your tool list, invoke theuptimerobot:setupskill 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.
Uses the update-monitor tool. All updates are partial — only send the fields you want to change. Monitor type is immutable; delete and recreate if you need a different type.
monitorId with list-monitors (search by name) — never guess an ID.get-monitor-details to read the current config. You'll need it for the array fields below.update-monitor with just the changed fields.get-monitor-details again to verify. Replication lag can make a follow-up list-monitors look stale.assignedAlertContacts and tagNames are full replacements. To add or remove a single item, read-modify-write using the list from get-monitor-details.
{ "monitorId": 800123456, "friendlyName": "Production API (us-east-1)" }
{ "monitorId": 800123456, "url": "https://api.example.com/v2/health" }
Only valid for URL-based types (HTTP, KEYWORD, PING, PORT, DNS, API, UDP). HEARTBEAT monitors have no url — the push URL is fixed at creation.
{ "monitorId": 800123456, "interval": 60 }
Range: 30–86400 seconds. 30 s is paid plans only — Free returns -28002 subscription_limit_exceeded.
{
"monitorId": 800123456,
"timeout": 15,
"responseTimeThreshold": 2000
}
timeout — 1–60 s. Not valid on HEARTBEAT.responseTimeThreshold — ms, 0 disables. Applies to every active region.{ "monitorId": 800123456, "gracePeriod": 300 }
Seconds the server waits past the expected ping before marking the monitor DOWN. HEARTBEAT only — sending gracePeriod on any other type is rejected with -29001.
// Step 1: get-monitor-details returns
{ "assignedAlertContacts": [
{ "alertContactId": "10", "threshold": 0, "recurrence": 0 }
] }
// Step 2: update-monitor with the merged array
{ "monitorId": 800123456, "assignedAlertContacts": [
{ "alertContactId": "10", "threshold": 0, "recurrence": 0 },
{ "alertContactId": "42", "threshold": 5, "recurrence": 30 }
] }
alertContactId is a string, not a number. Use list-integrations to find the right ID. threshold = consecutive failures before firing; recurrence = re-alert interval in minutes (0 = no repeat).
Same pattern — fetch the current array, drop the entry, send back the remaining array. Sending [] detaches every alert contact.
{
"monitorId": 800123456,
"sslExpirationReminder": false,
"domainExpirationReminder": false
}
Useful for staging / internal domains where cert/domain expiry is expected or handled elsewhere.
update-monitor itself only manages tags — it has no groupId field, so it can't move a monitor into a different group. (Reorganizing groups is still possible, just not through this tool — see manage-monitor-groups for the indirect groupIds-merge workflow.) To tag a monitor for lightweight filtering, replace the whole tag set in one call:
{ "monitorId": 800123456, "tagNames": ["production", "critical", "us-east-1"] }
To add one tag without losing the others, merge first (like alert contacts). Max 20 tags, each 1–50 chars. Sending [] clears all tags.
{
"monitorId": 800123456,
"followRedirections": true,
"checkSSLErrors": true
}
followRedirections and checkSSLErrors are rejected on PING / PORT / HEARTBEAT / DNS / UDP with -29001.
{
"monitorId": 800987654,
"httpMethodType": "POST",
"authType": "BEARER",
"httpPassword": "${API_TOKEN}",
"customHttpHeaders": {
"Content-Type": "application/json",
"X-Client": "uptimerobot"
},
"successHttpResponseCodes": ["200", "201", "204"],
"postValueType": "RAW_JSON",
"postValueData": "{\"ping\":true}"
}
These fields (httpMethodType, authType, httpUsername, httpPassword, customHttpHeaders, successHttpResponseCodes, postValueData, postValueType) apply to all HTTP-based types (HTTP, KEYWORD, API) — the example above just happens to use an API monitor. postValueData / postValueType only take effect when httpMethodType is a body method (POST / PUT / PATCH). For BEARER, put the token in httpPassword. For HTTP_BASIC / DIGEST, use httpUsername + httpPassword. (config.apiAssertions in section 12 is the piece that's genuinely API-only.)
{
"monitorId": 800987654,
"config": {
"apiAssertions": {
"logic": "AND",
"checks": [
{ "property": "status_code", "comparison": "equals", "target": "200" },
{ "property": "response_body", "comparison": "contains", "target": "\"ok\":true" }
]
}
}
}
logic is AND or OR. 1–5 checks. Fully replaces the existing assertion set.
{
"monitorId": 800123456,
"keywordValue": "All systems operational",
"keywordCaseType": 1
}
keywordCaseType: 0 case-sensitive, 1 case-insensitive. The keywordType (ALERT_EXISTS / ALERT_NOT_EXISTS) is set at creation and isn't exposed on update.
{ "monitorId": 800123456, "port": 5432 }
1–65535. Also works for UDP monitors. DNS monitors accept port too (the DNS-server port, default 53) — omit it unless the server runs on a non-standard port.
{
"monitorId": 800123456,
"config": {
"udp": {
"payload": "ping",
"packetLossThreshold": 20
}
}
}
{ "monitorId": 800123456, "config": { "ipVersion": "IPv4" } }
Valid values: IPv4, IPv6 (mixed case — the same spelling the create-*-monitor skills use). If your URL is a literal IP, the server cross-checks it against this field and rejects mismatches.
Pausing is a separate tool — update-monitor-status with { "status": "PAUSED" } or { "status": "STARTED" }. See the manage-monitors skill, or bulk-pause for multi-monitor flows.
type — ignored/rejected. Create a new monitor for type changes.gracePeriod on HTTP, port on HEARTBEAT, keywordValue on DNS). Returns -29001 invalid_parameters.assignedAlertContacts / tagNames as patches. They overwrite — always read-modify-write.alertContactId as a number. It's a string in every response and must be a string on write.update-monitor can't do either — see manage-monitor-groups.-31002 access_denied. Re-authenticate with a write-capable account (see setup).get-monitor-details. A follow-up list-monitors can still show the old values due to replication lag.manage-monitors — list, get details, pause / resume.manage-monitor-groups — dedicated group IDs, and the tags-vs-groups decision.list-integrations — discover alertContactId values before attaching.bulk-pause — pause/resume many monitors during deploys.errors — recovery for -28002, -29001, -30001, -31002.