An UptimeRobot MCP for workflows that run when nobody is watching
Connect one UptimeRobot API token and 56 nodes land on the canvas: create and update monitors across every check type, pause and start them in bulk around a deploy, read uptime and response-time statistics, work incidents and their comments, schedule maintenance windows, manage alert contacts and channel integrations, and post status page announcements. An AI agent can answer what is down and why, and a scheduled workflow can act on it without anyone in the room.
What the UptimeRobot MCP connector does
The UptimeRobot connector for Mattering Automations is 56 nodes on UptimeRobot API v3, 26 that read and 30 that act. Monitors are the centre of it: list, get, create, update and delete across the check types v3 supports, plus pause, start, reset statistics, and bulk pause, start and update by group or tag. Around them sit uptime statistics per monitor and account-wide, response-time series per monitor and per region, incidents with their activity log, sent alerts and comments, maintenance windows, personal alert contacts and channel integrations, monitor groups, tags, and public status pages with their announcements including pin and unpin.
Every node speaks MCP, the open protocol for connecting AI assistants to tools, so Claude, ChatGPT, and any MCP-capable assistant can use them, alongside Matter, the AI assistant built into the platform. UptimeRobot publishes an MCP server of its own, hosted by UptimeRobot, and it is a good one: it is client-agnostic, it takes an API token or an OAuth sign-in with a read-only or full-access choice, and it exposes monitors, recent incidents, integrations, monitor create and update, uptime figures, and response-time series with min, max, average and p95. If what you want is to ask questions about your monitoring from a chat client, use it. This connector is a different shape of the same access, and the differences are worth stating plainly rather than pretending they do not overlap.
What a connector adds is that the nodes are not only tools for an assistant, they are steps in a workflow. They run on a schedule or off a trigger with nobody present. They chain with every other connector in the same automation, so a monitor going down can open a task, post to a channel, write a row, and create a maintenance window in one run. The token is held once in a shared workspace credential store rather than in each person's own client config, so an agency or an on-call team is not distributing an account token by hand. And the destructive nodes carry flags, so a delete can be put behind a condition or an approval rather than sitting one sentence away from an agent.
Nine of the 56 nodes are destructive, and the ones worth naming are UptimeRobot: Delete Monitor, which takes the monitor and its entire uptime history with it, UptimeRobot: Reset Monitor Stats, which clears the statistics without deleting the monitor, and the deletes for alert contacts, integrations, maintenance windows, monitor groups, tags, status pages and incident comments. There is no trash and no undo behind any of them. One quieter risk is not a delete at all: pausing a monitor stops it being CHECKED, which means it stops ALERTING, so a deploy automation that pauses in bulk and then fails before the start step leaves an account silently unmonitored. The bulk pause and bulk start nodes are designed to be used together for exactly that reason.
One connection, three surfaces
- Visual automations. Drag UptimeRobot nodes onto the canvas and wire them to triggers, conditions, and other apps.
- AI agents, yours or ours. These nodes speak MCP, the open protocol, so Claude, ChatGPT, and any MCP-capable assistant can use them, and Matter, the agent built into the platform, uses them natively.
- The whole platform. Combine with native tasks, projects, and other connectors, no middleware in between.
Every UptimeRobot node, ready to drop on the canvas
56 nodes, generated from the connector itself so this list always matches what ships. Expand a category to browse.
Read and query nodes 26
UptimeRobot: Get Account
Get the account behind the API token: name email how many monitors exist how many the plan allows remaining SMS credits and the active subscription. Use monitorsCount against monitorLimit before a create-monitor step so a workflow can stop rather than fail on a plan limit.
UptimeRobot: Get Account Uptime Stats
Account-wide uptime across every monitor for a timeframe plus a log of the downtime and slow-response entries behind it. This is the closest thing v3 has to the old monitor-logs call: the logs array carries each entry's type timestamp and duration. Pick a preset timeframe or send CUSTOM with a start and end Unix timestamp. Note that overall uptime is a percentage across the whole window so it cannot be added up across windows.
UptimeRobot: Get Alert Contact
Get one personal alert contact with its type destination value which monitor events it is notified for whether SSL expiry reminders are on and whether it is active. A contact that is not Active has not completed verification and will not receive anything.
UptimeRobot: Get Incident
Get one incident with its root-cause detail: what failed the reason text the status the start and resolve times and the duration. Use it after UptimeRobot: List Incidents when you need the diagnosis rather than the summary row.
UptimeRobot: Get Incident Activity Log
The full timeline of one incident newest first: every status change with the region and checking node that observed it every comment and every notification. Rows differ by their type field so a STATUS_UPDATE row carries region and responseTime while a COMMENT row carries the comment. This is what to paste into a postmortem.
UptimeRobot: Get Incident Sent Alerts
Every alert UptimeRobot sent for one incident: who it went to which channel it used when it left and whether delivery succeeded. A NOT_DELIVERED row is the answer to why nobody reacted so this is the node an escalation workflow checks before assuming the team was told.
UptimeRobot: Get Integration
Get one notification channel with its type destination custom values and status. The meaning of the customValue fields depends on the channel type - a Slack integration puts the target channel there while a webhook uses them for its payload settings.
UptimeRobot: Get Maintenance Window
Get one maintenance window: its repeat interval start date and time duration in minutes which days it runs on whether new monitors are added automatically and which monitors it currently covers.
UptimeRobot: Get Monitor
Get one monitor's full configuration and current state: status check interval timeout regions assigned alert contacts tags custom fields SSL and domain expiry dates and the last 24 hours of uptime. Use this to read a monitor before updating it so an update only changes what you meant to change.
UptimeRobot: Get Monitor Group
Get one monitor group by id with its name and timestamps. To see which monitors are in it use UptimeRobot: List Monitors with the same group_id filter - the group record itself does not carry its members.
UptimeRobot: Get Monitor Response Times
Minimum maximum and average response time in milliseconds for one monitor over a date range plus how many measurements those figures came from. Defaults to the last 24 hours and the API caps a single request at 90 days. Turn on include_time_series to also get every individual measurement with its timestamp which is what a chart needs. Optionally narrow to one monitoring region.
UptimeRobot: Get Monitor Response Times By Region
The same response-time figures split by monitoring region in one call: separate min max average and measurement count under na eu as and oc. Use it to see whether a slowdown is global or regional. The payload is keyed BY REGION rather than being a list so downstream paths name a region directly - data.na.summary.avg - and a region the monitor does not check comes back null.
UptimeRobot: Get Monitor Uptime Stats
Uptime percentage total downtime incident count and mean time between failures for one monitor over a date range. Defaults to the last 24 hours and the API caps a single request at 90 days. The answer describes the WHOLE window as one figure: there are no per-day rows so the uptime percentage cannot be added up or averaged across windows without a denominator the API does not return. Read it per window you actually want to report on.
UptimeRobot: Get Status Page
Get one public status page with its url key custom domain covered monitor ids and tag ids pinned announcement and the full custom-settings block covering theme colours fonts and which widgets are shown.
UptimeRobot: Get Status Page Announcement
Get one status page announcement with its title body type start and end dates publication status subscriber delivery status and sent count.
UptimeRobot: List Alert Contacts
List the personal alert contacts the API token's own user owns: email addresses Pro SMS numbers voice calls and mobile push devices with each one's type value and activation status. Channel integrations like Slack and webhooks live separately under UptimeRobot: List Integrations. This is the node that populates alert-contact pickers.
UptimeRobot: List All Alert Contacts
Every alert contact reachable from this account grouped by the user who owns it: the token holder's own contacts notify-only contacts and each active organization member's contacts. Use it to audit who gets paged across a team. Note the response is a BARE ARRAY so rows are at data[] rather than data.data[] and each row nests its contacts under alertContacts.
UptimeRobot: List Incident Comments
Comments on one incident oldest first with who wrote each one and whether it is published on the public status page. Use it to build an incident timeline or to find the comment id an update or delete needs.
UptimeRobot: List Incidents
List downtime and slow-response incidents across the account with the monitor that raised each one why it fired when it started and how long it lasted. Filter by monitor by name or by start date range. An incident that is still ongoing has a null resolvedAt and a null duration - both fill in when it resolves - so a workflow polling for open incidents should key on resolvedAt being null rather than on status text.
UptimeRobot: List Integrations
List the notification channels wired to this account: Slack Discord Telegram Microsoft Teams Google Chat Mattermost Splunk PagerDuty Pushover Pushbullet Zapier and plain webhooks. Each row carries its id friendly name type destination and whether it fires on up down or both. Integration ids are what assigned_alert_contacts on a monitor accepts alongside personal contact ids.
UptimeRobot: List Maintenance Windows
List the account's maintenance windows with their schedule duration status and the monitors each one covers. Alerts are suppressed for those monitors while an active window is running so this is where to look when an expected alert never arrived.
UptimeRobot: List Monitor Groups
List the account's monitor groups with their ids names and timestamps. Groups are how the bulk pause start and update nodes select what to act on. Note the built-in default group for ungrouped monitors is id 0 and is NOT returned here - send 0 explicitly when a node needs it.
UptimeRobot: List Monitors
List the monitors on the account with their current status interval and configuration. Filters combine with AND logic. This is the node every other UptimeRobot node leans on: monitor pickers are populated from it and it is the cheapest way to confirm a credential works.
UptimeRobot: List Status Page Announcements
List a status page's announcements newest first with each one's title body type schedule publication status and how many subscribers it reached. Filter by status to separate drafts from what customers can actually see right now.
UptimeRobot: List Status Pages
List the account's public status pages with each one's url key custom domain covered monitors password state and styling. The public address is built from the url key so this is how an automation finds the link to share during an incident.
UptimeRobot: List Tags
List the account's monitor tags with their ids and names sorted by id. Tags are what the bulk nodes filter on and what the monitor lister's tags filter matches by NAME. Note this endpoint pages with nextCursorId rather than the nextLink the other list endpoints return.
Action nodes 30
UptimeRobot: Create Alert Contact
Create a personal alert contact. Email and mobile push are the types this endpoint can create; Pro SMS and voice need phone verification in the dashboard and are rejected here. A new email contact is not usable until the recipient confirms it so expect the status to come back as something other than Active and do not assign it to a monitor and assume alerts will arrive.
UptimeRobot: Create Incident Comment
Add a comment to an incident and optionally publish it on the public status page. This is how an automation narrates an outage: post what is known when the incident opens then post the resolution when the monitor recovers. Publishing puts the text in front of customers so treat that flag as a public statement.
UptimeRobot: Create Integration
Wire up a notification channel. The type picks the shape of the data object so a Slack integration wants a webhook URL and a channel while a plain Webhook integration wants a URL and a payload template. The new integration's id is what assigned_alert_contacts on a monitor accepts. Nothing is verified at creation time so a wrong webhook URL fails silently at the first alert - test it with a monitor you can safely pause and start.
UptimeRobot: Create Maintenance Window
Schedule a maintenance window that suppresses alerts for the monitors it covers. Choose a one-off or a daily weekly or monthly repeat then give it a start date a start time and a duration in minutes. Weekly and monthly repeats also need the days list. Creating the window does not pause the monitors: they keep being checked and the history stays continuous - only the alerting is held back.
UptimeRobot: Create Monitor
Create a monitor of any of the nine supported types. The type field decides which of the other fields apply: HTTP and KEYWORD check a URL, PING and PORT check a host, HEARTBEAT waits for an inbound check-in, DNS compares records, API asserts on a JSON response, UDP sends a payload, and VISUAL_COMPARISON watches a page for visual change. How many monitors an account may hold and how low the interval may go are both capped by the UptimeRobot plan, so a rejected create is usually a plan limit rather than a bad request.
UptimeRobot: Create Monitor Group
Create a monitor group and optionally fill it straight away either by listing monitor ids or by naming existing groups whose monitors should MOVE into the new one. A monitor belongs to exactly one group so both forms move monitors rather than copying them. Monitor groups are documented as a plan-gated feature and the API answers 403 when the account does not have it.
UptimeRobot: Create Status Page Announcement
Post an announcement to a public status page: a maintenance notice an incident update or general information. Setting status to Published makes it CUSTOMER VISIBLE immediately and can send it to the page's subscribers so draft with Offline or schedule with Pending when the wording still needs a human. This is the outward-facing half of an incident automation - UptimeRobot: Create Incident Comment is the internal half.
UptimeRobot: Delete Alert Contact
DESTRUCTIVE. Removes a personal alert contact permanently and with it every assignment to a monitor - those monitors simply stop notifying this destination and nothing warns you. There is no trash and no undo and a recreated email contact has to be verified again before it works. To take a contact off call temporarily set is_active false with UptimeRobot: Update Alert Contact instead.
UptimeRobot: Delete Incident Comment
DESTRUCTIVE. Removes a comment from an incident permanently - UptimeRobot has no trash and no undo so the text is gone. If the comment was published on the status page it disappears from the public timeline too which can leave a customer-visible gap in the incident narrative. Prefer UptimeRobot: Update Incident Comment to correct a comment rather than deleting it.
UptimeRobot: Delete Integration
DESTRUCTIVE. Removes a notification channel permanently along with every monitor assignment that pointed at it. Those monitors stop notifying that channel immediately and nothing warns you - the failure looks like an outage nobody was told about. There is no trash and no undo and recreating the channel means re-assigning it to every monitor by hand.
UptimeRobot: Delete Maintenance Window
DESTRUCTIVE. Removes a maintenance window permanently - UptimeRobot has no trash so the schedule and its monitor assignments are gone and the id stops resolving anywhere it was referenced. Alerts that the window was suppressing start firing again straight away. To silence a window temporarily set its status to paused with UptimeRobot: Update Maintenance Window instead.
UptimeRobot: Delete Monitor
DESTRUCTIVE and irreversible. Deleting a monitor takes its entire history with it: every incident every response-time measurement and every uptime figure ever recorded for it. UptimeRobot documents no trash and no restore anywhere in v3 so there is no way to get any of it back. If you only want the checks to stop use UptimeRobot: Pause Monitor which keeps the history.
UptimeRobot: Delete Monitor Group
DESTRUCTIVE. Removes a monitor group permanently. The monitors themselves survive and are moved to the group you name or to the default group 0 if you name none - so no monitoring is lost - but the grouping is gone with no undo and anything that selected monitors by this group id stops matching. Bulk pause and start rules pointed at this group silently select nothing afterwards. The default group 0 cannot be deleted.
UptimeRobot: Delete Status Page
DESTRUCTIVE and PUBLICLY visible. Removes a status page permanently: its url key stops resolving so any link a customer or a vendor questionnaire has bookmarked goes dead and its announcement history goes with it. There is no trash and no undo. Note this connector cannot recreate the page for you - creating and updating a status page are multipart file uploads the canvas cannot produce so recovery means the dashboard.
UptimeRobot: Delete Tag
DESTRUCTIVE. Deletes a tag and strips it from EVERY monitor carrying it in one call. The monitors survive but their labelling is gone with no undo and anything selecting monitors by this tag - bulk pause rules status page filters saved searches - silently matches nothing afterwards. Check what would be affected with UptimeRobot: List Monitors filtered by the tag name first.
UptimeRobot: Pause Monitor
Stop checking a monitor until it is started again. Not destructive - the history is kept and the monitor resumes exactly as configured - but say it plainly: a paused monitor STOPS ALERTING so an outage while it is paused goes unreported. Idempotent: pausing an already paused monitor succeeds.
UptimeRobot: Pause Monitors In Bulk
Pause every monitor matching a group a tag or both. Supplying both narrows to monitors matching BOTH filters. At least one filter is required so there is no way to pause the whole account by accident. Not destructive but every matched monitor STOPS ALERTING until it is started again - the usual use is a maintenance window that a deploy pipeline opens and closes. The response reports success or error per monitor.
UptimeRobot: Pin Status Page Announcement
Pin an announcement to the top of a public status page so it is the first thing a visitor reads. Idempotent: pinning an already pinned announcement succeeds. A status page tracks ONE pinned announcement at data.pinnedAnnouncementId so pinning a second one replaces the first.
UptimeRobot: Reset Monitor Stats
DESTRUCTIVE and irreversible even though nothing is deleted. The monitor keeps running but its accumulated statistics go to zero and its incidents and sent alerts are cleared. Every uptime figure and every incident record for this monitor is gone with no way to restore it. Read the numbers you need with UptimeRobot: Get Monitor Uptime Stats BEFORE running this.
UptimeRobot: Start Monitor
Resume a paused monitor so checks and alerting start again. Idempotent: starting an already running monitor succeeds. Pair this with UptimeRobot: Pause Monitor around a deploy so the window is closed even if the deploy step fails.
UptimeRobot: Start Monitors In Bulk
Resume every paused monitor matching a group a tag or both. Supplying both narrows to monitors matching BOTH filters and at least one is required. This is the closing half of a deploy-window automation: pause in bulk before the deploy and start in bulk after it. The response reports success or error per monitor.
UptimeRobot: Unpin Status Page Announcement
Remove an announcement from the top of a public status page. The announcement itself is untouched and stays in the list. Idempotent: unpinning something that is not pinned succeeds. This is the closing step of an incident workflow - unpin once the monitor recovers.
UptimeRobot: Update Alert Contact
Rename a personal alert contact change which events it hears about toggle its SSL expiry reminders or pause it. Only the fields you fill are sent. The destination itself cannot be changed here: to point alerts at a different address create a new contact. Setting is_active to false silences the contact everywhere it is assigned without deleting it which is the safe way to take someone off call.
UptimeRobot: Update Incident Comment
Rewrite an existing incident comment or change whether it is published on the status page. The content field is required so send the full replacement text rather than an addition. Editing a published comment changes what customers see immediately.
UptimeRobot: Update Integration
Change an existing notification channel: rename it point it somewhere else or change which events it fires on. Send the same type the integration already has - the type selects which shape the data object is validated against and is not a way to convert a Slack integration into a webhook.
UptimeRobot: Update Maintenance Window
Change a maintenance window in place or pause and resume it. Only the fields you fill are sent. monitor_ids REPLACES the covered set rather than adding to it so send the full intended list. Setting status to paused stops the window SUPPRESSING alerts which is the opposite of pausing a monitor - a paused window means alerts get through again.
UptimeRobot: Update Monitor
Change an existing monitor in place. Only the fields you fill are sent so everything else is left alone. Two fields do NOT merge the way the rest do: assigned_alert_contacts and tag_names REPLACE the current lists rather than adding to them so send the full intended list. The config object is a shallow merge - omit a key to keep it and set a key to null to remove it.
UptimeRobot: Update Monitor Group
Rename a monitor group. Renaming is all this endpoint does - to move monitors between groups set each monitor's group_id with UptimeRobot: Update Monitor or UptimeRobot: Update Monitors In Bulk.
UptimeRobot: Update Monitors In Bulk
Apply the same settings change to every monitor matching a group a tag or both. Only the settings you fill are changed. At least one filter is required and supplying both narrows to monitors matching BOTH. The blast radius is whatever matches at that moment so check it with UptimeRobot: List Monitors using the same filter first. The response reports success or error per monitor.
UptimeRobot: Update Status Page Announcement
Edit an announcement or move it between states. Only the fields you fill are sent. Moving status from Offline to Published is how a draft goes live and moving it to Archived takes it off the page while keeping the history - both are immediately visible to customers. There is no delete endpoint for announcements in v3 so Archived is the intended way to retire one.
Connecting UptimeRobot takes one credential
UptimeRobot API v3 authenticates with a single token, sent as an Authorization Bearer header, and there is nothing else to configure: no account id, no version header, no second secret. Create it in the UptimeRobot dashboard under Integrations and API, and paste the raw token; the Bearer prefix is added for you, which matches UptimeRobot's own note that the token is entered without it. Build against v3, not the legacy v2 API. They are two different APIs wearing one brand: v2 is POST-only for every operation including reads, with the key in the request BODY and verb-shaped paths, so anything written from an old example or from memory will fail completely against v3. UptimeRobot's v3 announcement says v2 will remain available but will no longer receive support or updates and that new features go to v3 only, and it publishes no sunset date, so v2 is frozen rather than scheduled for removal. On plans: UptimeRobot's API page states the API is included across all plans, from Free to Enterprise, so a free account can connect this and run the read nodes. Some node families are gated, though, and the gating is documented in UptimeRobot's own API specification rather than only on marketing pages, which makes it worth stating exactly. Monitor groups are described as a feature that may not be available, and answer 403 when it is not. Custom fields on monitors are annotated as paid plans only. The threshold and recurrence settings on alert contacts are annotated as available only in the paid plans and always zero on the Free Plan. UptimeRobot's API page also states the free tier as up to 50 monitors with 5-minute monitoring intervals, which matters here because the check interval is a required field when a monitor is created and a value below your plan's floor is rejected by the API. Check UptimeRobot's own pricing page for what your plan includes before building a workflow around a gated family.
Sign in to UptimeRobot and open Integrations and API. Create an API token and copy it.
In Mattering Automations, open Connectors, click Add Connector, choose UptimeRobot, and create a credential with that token. Paste the raw value; the Bearer prefix is added for you.
Click Test Connection. Mattering calls UptimeRobot's monitor list endpoint, which needs no parameters and has no side effect, so a response, even an empty list, proves the token.
All 56 UptimeRobot nodes appear in the canvas palette. Run UptimeRobot: List Monitors first: it returns the monitor ids that eleven other nodes take, and the monitor picker on those nodes is filled from it, grouped by check type and described by the URL each monitor checks.
Common uses for the UptimeRobot MCP
Provision monitoring from wherever the truth already lives
UptimeRobot: Create Monitor puts a check behind anything you ship, and because it is a node it can run off the same event that created the thing being monitored: a new client record, a new deployment, a new subdomain, a row added to a Mattering database. UptimeRobot: Update Monitor keeps it in step, UptimeRobot: Create Monitor Group and UptimeRobot: List Tags organise the result, and UptimeRobot: Delete Monitor cleans up when a service is retired. Two things the nodes will tell you and a script will not. The check interval is required at creation and a value below your plan's floor is rejected, and UptimeRobot's API page states the free tier at 5-minute intervals. And on an update, the assigned alert contacts and the tag names REPLACE rather than add, so an update that sends one contact removes every other contact from that monitor.
Deploy windows that do not leave you blind
UptimeRobot: Pause Monitors In Bulk and UptimeRobot: Start Monitors In Bulk are built as a pair, and the reason is the failure they prevent. Pausing a monitor stops it being checked, which means it stops alerting, so a deploy workflow that pauses everything and then fails before it resumes leaves the account silently unmonitored. Put the start step on the failure path as well as the success path. Bulk operations also require at least one filter, a group or a tag, so there is no accidental all-monitors form, though the blast radius is still whatever matches at that moment. The alternative shape is UptimeRobot: Create Maintenance Window, which suppresses alerts on a schedule instead of stopping the checks. Note the word pause inverts between the two: a paused maintenance window stops SUPPRESSING, so alerts get through.
Incident triage that reaches the rest of your stack
UptimeRobot: List Incidents is the node a schedule polls, and an ongoing incident is identified by its resolution timestamp still being null rather than by any status text, which is what the node tells you to test on. From there UptimeRobot: Get Incident Activity Log and UptimeRobot: Get Incident Sent Alerts explain what happened and who was told, and UptimeRobot: Create Incident Comment writes the human context back so the timeline in UptimeRobot matches the one in your task tracker. Because this is a canvas, the same run can open a Mattering task, post to a channel, and pull the deploy history from another connector, so the person who picks it up gets the whole picture rather than a ping.
Status page announcements as a workflow step
UptimeRobot: Create Status Page Announcement, UptimeRobot: Update Status Page Announcement, UptimeRobot: Pin Status Page Announcement and UptimeRobot: Unpin Status Page Announcement let the customer-facing message be part of the incident workflow rather than a task somebody forgets under pressure. Draft it with an agent from the incident data, post it pinned when the incident opens, update it as things change, and unpin it on resolution. UptimeRobot: List Status Pages and UptimeRobot: Get Status Page find the page to post to. One honest boundary: creating and updating the status page ITSELF is not available here, because UptimeRobot declares those two operations as multipart form uploads only, which a workflow canvas cannot produce. Reading, listing and deleting status pages are all built, and the whole announcement surface underneath them is.
Alert routing kept in configuration rather than in someone's memory
UptimeRobot: List Alert Contacts, UptimeRobot: Create Alert Contact and UptimeRobot: Update Alert Contact manage who gets told, and UptimeRobot: List All Alert Contacts is the only way to see organization members' contacts rather than just your own. UptimeRobot: Create Integration and UptimeRobot: Update Integration manage the channel side, Slack, Discord, Teams, webhooks and the rest. Two silent failures the nodes warn about, both of which look fine until an outage. A new email alert contact is not usable until it is VERIFIED, so creating one and assigning it does not mean alerts will arrive. And integrations are not validated at creation, so a wrong webhook URL is accepted and fails at the first real alert rather than at setup. On the threshold and recurrence settings, UptimeRobot's API specification annotates them as paid-plan only and always zero on the Free Plan.
Uptime and response-time reporting, read correctly
UptimeRobot: Get Monitor Uptime Stats returns uptime, total downtime, incident count and mean time between failures for a window. UptimeRobot: Get Monitor Response Times returns the measurement series, and UptimeRobot: Get Monitor Response Times By Region splits it by checking location. UptimeRobot documents a maximum range of 90 days on these, so a longer history is several requests. One caution matters more than the rest, and it is the reason this connector deliberately declares no aggregation rules on these numbers: uptime is a PERCENTAGE, and UptimeRobot returns no denominator column beside it, so an average of daily uptime figures is not the month's uptime and nothing in the response lets you weight it correctly. Re-request the window you want to report on rather than recombining smaller ones. Downtime seconds and incident counts describe the window they were asked for rather than any row inside it, and response-time rows are individual measurements at the monitor's own check interval, which on a one-minute monitor over 30 days is tens of thousands of points in one execution.
Example: A deploy window that pauses monitoring and always resumes it
Built on the Automations canvas in a few minutes, or described in plain language to Matter and built for you.
Trigger: your deploy pipeline calls a Mattering webhook, or a scheduled release window opens.
UptimeRobot: Pause Monitors In Bulk pauses the monitors tagged for this service. Bulk operations require a group or a tag filter, so the scope is whatever that tag matches at that moment rather than everything.
UptimeRobot: Create Maintenance Window optionally covers the same period, for the checks you would rather suppress alerts on than stop entirely.
The deploy runs, and the workflow waits on its result.
UptimeRobot: Start Monitors In Bulk resumes the same tag. This step is on BOTH the success path and the failure path, because a paused monitor is not alerting, and a run that dies between the pause and the start leaves the service unmonitored with nothing to announce it.
UptimeRobot: List Incidents runs a few minutes later, filtered to incidents that started after the deploy began, and a condition tests for a resolution timestamp still being null rather than for status text.
If anything is still open, UptimeRobot: Create Incident Comment records the deploy reference on the incident, a Mattering task opens for whoever is on call with the incident and the deploy details attached, and UptimeRobot: Create Status Page Announcement posts a pinned note if the affected monitors are on a public status page.
UptimeRobot MCP questions, answered
What UptimeRobot operations are supported?
56 nodes on API v3: 26 read and 30 act. Monitors: list, get, create, update, delete, pause, start, reset statistics, and bulk pause, start and update by group or tag, plus uptime statistics per monitor and account-wide and response-time series per monitor and per region. Incidents: list, get, activity log, sent alerts, and comments with create, update and delete. Maintenance windows: full create, read, update and delete. Alert contacts: list personal and organization-wide, get, create, update, delete. Integrations: list, get, create, update, delete. Monitor groups and tags. Status pages: list, get and delete, plus the whole announcement surface with create, update, pin and unpin. Account details. Three documented operations are deliberately not shipped: creating and updating a status page, because UptimeRobot declares both as multipart form uploads only and a workflow canvas cannot produce one, and a near-duplicate personal alert contact listing that returns a worse shape than the one that is built.
Does this work with Claude, ChatGPT, or other AI assistants?
Yes. Mattering exposes these nodes over MCP, an open protocol for connecting AI assistants to tools, so any MCP-capable assistant, including Claude and ChatGPT, can connect to your workspace and use them. Matter, the assistant built into the platform, uses the same nodes natively. The difference from a chat-only setup is that the same nodes also run as steps on an automation canvas, on a schedule or off a trigger, with nobody present.
UptimeRobot has its own MCP server. Why use this connector?
It does, it is hosted by UptimeRobot so there is no local server to run, it is client-agnostic with ready-made setups for several editors, and it takes either an API token or an OAuth sign-in where you approve read-only or full access. It covers monitors and their status, recent incidents per monitor, connected integrations, creating and updating monitors across the check types, uptime figures, and response-time series with min, max, average and p95. For asking questions about your monitoring from a chat client, that is the direct route and worth using. This connector answers a different need. The nodes are workflow STEPS as well as assistant tools, so they run on a schedule or off a trigger with nobody in the room, and they chain with every other connector in one automation, which is what lets a deploy pause monitors, resume them on both the success and failure paths, open a task, and post a status page announcement in a single run. The credential lives once in a shared workspace store rather than in each person's client config, which matters for a team or an agency. The node set is wider, covering maintenance windows, alert contacts, monitor groups, tags, incident comments and the full status page announcement surface. And the nine destructive nodes are flagged so you can gate them behind a condition or an approval. The two are not exclusive: use UptimeRobot's server for interactive questions and this connector for the work that has to happen unattended.
Does it use the v2 or the v3 UptimeRobot API?
v3, exclusively, and the difference is not cosmetic. v2 is POST-only for every operation including reads, puts the key in the request BODY, and uses verb-shaped paths. v3 uses ordinary REST verbs, resource paths, cursor pagination and an Authorization Bearer header. Anything written from an old integration or from memory will be v2-shaped and will fail completely. UptimeRobot's v3 announcement says v2 remains available but will not receive support or updates and that new features go to v3 only; no sunset date is published, so v2 is frozen rather than retiring on a date. One dated change does live inside v3's payloads: the mobile push alert-contact type names change to their canonical spellings on 10 October 2026, and UptimeRobot tells consumers to accept both forms in the meantime. Nothing in this connector branches on that value, so the switchover is a non-event here.
Do I need a paid UptimeRobot plan?
Not to connect. UptimeRobot's API page states the API is included across all plans, from Free to Enterprise, with no separate API charge, so a free account can create a token and run the read nodes. Some families are gated, and because the gating is written into UptimeRobot's own API specification rather than only into marketing copy, it can be stated precisely. Monitor groups are described as a feature that may not be available and answer 403 when it is not, which reads as a permission error rather than a plan one. Custom fields on monitors are annotated as paid plans only. The threshold and recurrence settings on alert contacts are annotated as available only in the paid plans and always zero on the Free Plan. UptimeRobot's API page also states the free tier as up to 50 monitors with 5-minute monitoring intervals, which matters when you create a monitor because the interval is required and a value below your plan's floor is rejected. Confirm the current details on UptimeRobot's own pricing page before you build a workflow around a gated family.
What are UptimeRobot's rate limits?
UptimeRobot's API page states 10 requests per minute on the Free plan, rising with the plan up to 5,000 requests per minute on Pro, and its v3 announcement describes the paid figure as derived from your monitor limit. Ten a minute is low enough to shape how you build. A configuration panel with several pickers open, a scheduled refresh, and a workflow polling incidents can exceed it on a free account without anyone doing anything unusual. Option pickers are served from a short-lived cache, which is the main protection, but anything looping per-monitor statistics calls should pace itself. v3 standardises 429 responses and several endpoints list them explicitly, so a rate-limited call fails visibly rather than silently.
Can this trigger a workflow when a monitor goes down?
Not directly, and it is better to say so than to imply an event stream. This connector ships no trigger nodes, because UptimeRobot's own API has no subscription for us to expose. Two honest routes get you there. The first is polling: run UptimeRobot: List Incidents on a schedule and test for a resolution timestamp that is still null, which is what the node tells you to test on rather than status text, since an ongoing incident carries a null resolution time and a null duration that both fill in later. The second is UptimeRobot's own alerting: create a webhook integration with UptimeRobot: Create Integration pointing at a Mattering webhook trigger, so UptimeRobot pushes the alert to your workflow. Remember that integrations are not validated at creation, so a wrong URL is accepted and fails at the first real alert; test it deliberately rather than assuming.
Can an AI agent take my monitoring down by accident?
Two categories are worth gating, and the second is the one people miss. Nine nodes are destructive with no trash and no undo: delete monitor (which takes its entire uptime history with it), reset monitor statistics, and the deletes for alert contacts, integrations, maintenance windows, monitor groups, tags, status pages and incident comments. The quieter risk is UptimeRobot: Pause Monitor and UptimeRobot: Pause Monitors In Bulk, which are not destructive by any flag but stop the monitor being checked, and therefore stop it alerting, so a paused monitor is an unmonitored service that looks fine on a dashboard. Always pair a pause with a start on both the success and failure paths. Two update behaviours deserve the same care: the assigned alert contacts and tag names on a monitor update REPLACE rather than add, so a partial update silently removes routing, and the configuration object on an update is a shallow merge where sending an explicit null clears everything. Keep agents on the read nodes, gate the writes behind a condition or an approval task.
Why does my uptime chart disagree with UptimeRobot's own number?
Usually because it was assembled from parts. Uptime is a percentage, and UptimeRobot returns no denominator column beside it, no checks performed and no seconds observed, so there is no correct way to average daily figures into a monthly one. This connector deliberately declares no aggregation rule for it, which means a well-behaved consumer refuses to add it up rather than doing so wrongly. Ask UptimeRobot: Get Monitor Uptime Stats for the window you actually want to report on. The same applies to the window totals beside it: downtime seconds and incident counts describe the whole requested range rather than any row inside it. Two more sources of disagreement worth knowing: UptimeRobot documents a maximum range of 90 days on these endpoints, so a year is several requests, and the account-wide uptime node takes a time-frame mode where the start and end dates only apply when the mode is set to custom.
Put UptimeRobot to work inside your workspace
Connect once, then use every node in visual automations and AI agent workflows across the Mattering platform.
All product names, logos, and brands are property of their respective owners and are used for identification purposes only. Use of these names, logos, and brands does not imply endorsement or affiliation. UptimeRobot is a product of its respective owner; Mattering is an independent platform.