An Instatus MCP that knows which nodes your customers can see
109 nodes across status pages, components, incidents and their timelines, maintenance windows, saved templates, general notices, recorded outages, subscribers, metrics, workspaces and teammates, uptime monitors, cron monitors, monitoring integrations, routing rules, escalation policies and on-call schedules. 33 read and 76 act. The ones that publish to a live page or mail a subscriber list are flagged as such, because an automation that opens an incident is telling your customers something is broken.
What the Instatus MCP connector does
The Instatus connector for Mattering Automations is 109 nodes on the Instatus API. Status pages, components and their five settable statuses cover the page itself. Incidents, incident updates, saved templates, maintenance windows, maintenance updates, general notices and recorded outages cover everything a visitor reads. Subscribers, audience groups and teammates cover who sees it and who gets told. Metrics and their data points cover the public charts. Uptime monitors, cron monitors, monitor alerts, monitor groups and third-party monitoring integrations cover the detection side, and routing rules, escalation policies and on-call schedules cover who gets paged when detection fires. 33 nodes read and 76 act.
Every capability is a node on the Automations canvas, and every node speaks MCP, the open protocol for connecting AI assistants to tools. Claude, ChatGPT, and any MCP-capable assistant can use them, alongside Matter, the AI assistant built into the platform. The same nodes are steps in a visual workflow, so an incident response an assistant helps you design can then run on a schedule or off another system's alert with nobody present.
Read this part before you wire anything to a trigger, because it is the thing that makes an Instatus connector different from a database connector. An Instatus status page is a public web page, usually linked from a product's footer, and its subscribers receive notifications. So a node here is not writing to a private record: Instatus: Create Incident publishes to the live page by default, which is Instatus's own default rather than ours, and any node with notify turned on messages every confirmed subscriber of the affected components. An automation that creates an incident is an automation that tells your customers something is broken. That can be exactly what you want at three in the morning. It is a bad thing to discover by accident during a test run.
The connector marks that risk rather than leaving you to infer it. 38 of the 109 nodes carry a destructive flag, and only 23 of those delete anything. The other 15 delete nothing at all: they publish to a live page, mail a real person, or invalidate a link somebody is currently using. Instatus: Create Incident, Instatus: Schedule Maintenance, Instatus: Add Incident Update, Instatus: Add Teammate, Instatus: Add Subscribers In Bulk and Instatus: Regenerate Secure Link are all in that second group. Every one of them says in its own description what it does to a customer, so the warning arrives while you are configuring the node rather than afterwards.
Most paths in this API are scoped to a status page, so page_id is a first-class parameter rather than an afterthought: 83 of the 109 nodes take one, and every one of those fields is a picker fed by Instatus: List Status Pages, labelled with the page name and its subdomain. The child pickers for components, incidents, maintenances, templates, metrics, subscribers, audience groups, monitors, cron monitors and escalation policies fill in once a page is chosen and refill when it changes. The id Instatus wants is what its dashboard calls the Project ID, not the subdomain, which is the single most common thing to get wrong by hand and the reason the pickers exist.
One node needs no credential at all. Instatus: Get Public Page Summary reads the summary any Instatus status page publishes at its own address, so it works against any Instatus page including one belonging to a vendor you have no relationship with. That makes supplier monitoring a real workflow here rather than a stretch: watch the status pages your product depends on, and open a Productivity task or post to a channel the moment one of them stops reading UP. It uses the condensed three-value vocabulary of UP, HASISSUES and UNDERMAINTENANCE rather than the longer list the authenticated API returns.
Instatus does not gate the API behind a paid plan. Its pricing page says of API access, in its own words, that it is available on all plans, and the Starter plan is free forever with no card, so you can create a key and run this connector without paying anything. Two behaviours are plan-limited and both are named on the nodes that touch them: the WELCOME and NONE subscriber confirmation modes are Pro only, as is the autoConfirm override on the two subscriber-create nodes. Private and select-audience status pages are an upgrade beyond Starter, which is what the audience-group nodes and Instatus: Regenerate Secure Link depend on.
One connection, three surfaces
- Visual automations. Drag Instatus 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 Instatus node, ready to drop on the canvas
109 nodes, generated from the connector itself so this list always matches what ships. Expand a category to browse.
Read and query nodes 33
Instatus: Get Audience Group
Get one audience group with its teammates fully expanded to their user records, and its components expanded to the full component objects. Read it before updating, because the update endpoint REPLACES the lists it is given rather than adding to them.
Instatus: Get Component
Get one component in full: name, description, current status, order, group, uptime-bar visibility and its unique ingest email address. Read a component before updating it so an update changes only what you meant to change, since the update endpoint takes the same field set and an omitted field is not the same as an unchanged one for every field.
Instatus: Get Cron Monitor Summary
Get a cron monitor's reliability at a glance: total runs, failed runs, overall availability, and one, seven and thirty day availability each paired with the same figure for the previous period so a trend is readable without arithmetic. Also the last success, last failure, last start and how long it has been up or down. These windows are fixed and trailing, with no parameters at all, so this node reports a moving picture and cannot be asked about a historical range; for that, read the logs. The availability figures are percentages with no denominator in the payload, so read and compare them rather than averaging them together.
Instatus: Get General Notice
Get one general notice with its name, message in plain text and HTML, notify flag, collapsed flag and timestamps. Read it before updating, since the update endpoint applies only the fields you send.
Instatus: Get Incident
Get one incident with its complete update timeline, the components it affects and their statuses during it, and who raised it. The updates array is the incident's public history in order, so this is the node to read when summarising an outage or when deciding which status transition to post next.
Instatus: Get Incident Update
Get one incident update: its message in both plain text and rendered HTML, the status it set, whether it notified subscribers, when it started, and a summary of the incident it belongs to. There is no endpoint that LISTS incident updates on their own, so the ids for this node come from an incident's updates array via Instatus: Get Incident.
Instatus: Get Maintenance
Get one maintenance window in full: schedule, duration, current status, every update posted on it, the components it covers and all four notification flags. Read it before updating so a change to the schedule does not silently reset the notification settings, and read it before posting a transition so you do not duplicate one autoStart already made.
Instatus: Get Maintenance Update
Get one maintenance update: its message in plain text and HTML, the status it set, whether it notified subscribers, when it took effect, and a summary of the maintenance window it belongs to. As with incident updates there is no lister, so the ids come from the window's updates array via Instatus: Get Maintenance.
Instatus: Get Metric
Get one metric with its name, unit suffix, order, active flag and its data points as timestamp and value pairs. Timestamps are epoch MILLISECONDS, not seconds and not ISO strings, which is the same encoding the write endpoints expect. There is no way to request a date range, so this returns the window Instatus chooses.
Instatus: Get Outage
Get one recorded outage period with its start and end, severity, visibility state, the component it belongs to and the incident that produced it. Useful when reconciling a reported availability figure against the specific periods that produced it.
Instatus: Get Public Page Summary
Read the public summary any Instatus status page publishes at its own address, with the page's overall status plus its active incidents and active maintenances. Two things make this node different from everything else here. It talks to the status page's own host rather than to the API, and it needs no credential at all, so it works against ANY Instatus page including a vendor's, which makes it the node for watching a supplier you depend on. It also uses the condensed three-value status vocabulary of UP, HASISSUES and UNDERMAINTENANCE rather than the seventeen-value list the authenticated API returns.
Instatus: Get Template
Get one template in full: its type, name, message in plain text and HTML, the status it posts, whether it notifies subscribers and which components it changes. Worth reading before an automation posts it, because the template's own notify flag is what decides whether that automation emails your customers.
Instatus: Get User Profile
Get the Instatus user the API key belongs to: id, email, name, slug and avatar. An Instatus key is issued to a person rather than to an organisation, so this is how a workflow answers whose key it is running as, and how an audit trail records who an automated incident was posted by.
Instatus: List Audience Groups
List a private status page's audience groups, each with its name, the teammates in it and the components those teammates can see. Audience groups are how one private page shows different services to different customers, so this is the node for auditing who sees what. The path uses a hyphen: audience-groups on reads, audience_groups with an underscore on writes, which is Instatus's own inconsistency.
Instatus: List Components
List the components on a status page with each one's id, name, description, current status, display order, group and whether it shows an uptime bar. Components are the individual services the page reports on, and their ids are what incidents, maintenances, subscriptions and audience groups attach to, so this is the lookup most other nodes in this connector depend on. Each component also carries a uniqueEmail address: sending mail to it raises an incident on that component, which is how Instatus ingests alerts from tools with no integration.
Instatus: List Cron Monitor Logs
Read a cron monitor's check-in history: each ping with its status, the HTTP method used, the user agent and source address, when it arrived and when the job said it started. The status vocabulary tells you what actually happened: SUCCESS and FAILURE are pings the job sent, MISSED means nothing arrived inside the window, LATE means it arrived after the period but inside the grace, and STARTED is the opening half of a paired start and finish ping. Filter to important rows only, and bound the range with ISO 8601 dates; Instatus defaults to the last year. The limit here caps at 127 and the page number at 1000, which are the vendor's own numbers.
Instatus: List Cron Monitors
List the cron monitors on a status page: scheduled jobs that prove they ran by pinging Instatus, so a job that stops running is caught by its silence rather than by a failed request. Each row carries its expected period and grace in SECONDS, its current status, its state of ACTIVE, PAUSED or MUTED, the component it reports against, its ping slug and its full onFail and onRecover behaviour. Filter by status to find jobs that have missed their window.
Instatus: List Escalation Policies
List a status page's escalation policies with their ordered rules, each rule's delay in minutes, its condition and the actions it takes, plus the policy-level repeat settings. An escalation policy is the answer to nobody picked up: rule one pages the on-call engineer, rule two waits ten minutes and pages their backup, and repeat decides how many times the whole ladder runs before autoResolveIncidentsAfterRepeat gives up. Takes no pagination parameters.
Instatus: List General Notices
List the general notices on a status page. A general notice is a persistent banner that is not an incident and not a maintenance, which makes it the right shape for a standing announcement such as a migration window, a deprecation or holiday support hours. Each one carries its name, message in plain text and HTML, whether it notified subscribers and whether it renders collapsed.
Instatus: List Incidents
List a status page's incidents with each one's name, opening and resolution timestamps, current status, full update timeline, affected components and the workspace member who created it. Filter to or away from particular statuses with the status and exclude_status fields, so a workflow can ask for open incidents only without reading the whole history. The user field is null when the incident was created automatically by a monitor, a third-party integration or an inbound email, which is how you tell machine-raised incidents from human ones.
Instatus: List Maintenances
List a status page's maintenance windows with each one's name, scheduled start, duration in minutes, current status, its update timeline and its notification settings. The notifyEarly and notifyMinutes pair is the advance warning subscribers get; autoStart and autoEnd say whether Instatus moves the window through its lifecycle on the schedule by itself. Listed on /v2, read individually on /v1.
Instatus: List Metrics
List the metrics charted on a status page, each with its name, unit suffix, display order, active flag and its recorded data points as timestamp and value pairs. Note that the data points come back inside the metric rather than from a separate endpoint, and there is no date-range parameter anywhere on this API, so you get whatever window Instatus decides to return and cannot ask for a different one.
Instatus: List Monitor Alerts
List the monitor alerts configured on a status page, each with its channel, recipient, which state changes it fires on, its SSL expiry warning window and the monitors attached to it. An alert is the delivery half of monitoring: a monitor decides something is wrong and an alert decides who hears about it and how. Read this to audit who is being paged. Rows arrive under a monitorAlerts key with total, page, totalPages and limit alongside.
Instatus: List Monitor Logs
Read the individual check results for one monitor: when each ran, from where, whether it succeeded, the HTTP status it saw, and the timing breakdown of DNS, TCP, TLS, first byte, download and total response time. This is the node for latency investigation and for proving what actually happened during an outage. Filter by date range, location, success, status or HTTP code. Read the timings as a series rather than adding them up: each row is one measurement, so summing response times means nothing and averaging them across days is only honest when each day contributed the same number of checks.
Instatus: List Monitors
List the uptime monitors on a status page with each one's target URL, type, assertions, check interval, monitoring locations, alerts and its onFail and onRecover behaviour. Filter by status to find what is currently broken. This lister populates every monitor picker in the connector. Note the response shape and the pagination differ from the versioned half of this API: rows arrive under a monitors key alongside total, page, totalPages and limit, and the page size parameter is limit rather than per_page.
Instatus: List On-call Schedule Members
Read who is on call on a schedule right now, who is up next, and everyone in the rotation, each with their id, name, email, phone and avatar. This is the node that answers who do I contact, so it is the one worth wiring into a Matter agent or an incident-response workflow. Note that Instatus ships no endpoint that lists SCHEDULES, so the schedule id has to come from a create response or from the dashboard URL.
Instatus: List Outages
List the recorded outage periods on a status page: the from and to timestamps, the component affected, the severity, whether the period is published, and the incident it came from when there was one. Outages are the underlying record the uptime bars and uptime percentages are computed from, which makes this the node to read for real availability reporting rather than for the current status. An outage with a null incident was recorded directly rather than raised through the incident flow.
Instatus: List Routing Rules
List a status page's routing rules with their assertions and actions in order. A routing rule matches on an incoming event's title, description, impact or affected components and then notifies a team, a named workspace member, or whoever is currently on call. This is the node for auditing who gets paged for what, and the order field is significant. Takes no pagination parameters at all, unlike most of this API.
Instatus: List Status Pages
List the status pages your Instatus API key can see, with each one's id, name, subdomain, custom domain, current derived status and branding. This is the node the rest of the connector leans on: every page picker is populated from it, and it is the cheapest way to confirm a credential works. The id it returns is what Instatus calls the Project ID and what every other node here wants. Note the response is a bare JSON array with no total, so paginate by asking for the next page until one comes back empty. This list also carries the secure link for any private page you have member or owner access to.
Instatus: List Subscribers
List the people subscribed to a status page, with each one's email or phone number, webhook target, whether they have confirmed, and whether they follow the whole page or specific components. Search by email address or phone number to check whether a particular customer is subscribed. This payload is personal data: it is a customer contact list, so treat what a workflow does with it accordingly.
Instatus: List Teammates
List the people with access to a status page, each with their teammate id and their user record of name, email and avatar. The teammate id, not the user id, is what audience groups and the remove endpoint take, and the two are different values in the same payload, so read carefully. Note the path segment is team rather than teammates.
Instatus: List Templates
List the saved incident and maintenance templates on a status page, with each one's type, name, message, the status it posts, its notify flag and the component statuses it applies. The type field is INCIDENT or MAINTENANCE and it decides which endpoints will accept the template, so read it before wiring one into an automation. This node also populates every template picker in the connector.
Instatus: List Workspaces
List the Instatus workspaces your key belongs to, with each one's name, slug, the status page ids it contains, its SSO and directory-sync state, allowed email domains and the default notification settings applied to new members. A workspace is the billing and membership container above status pages, so this is where you look to find which pages sit together.
Action nodes 76
Instatus: Add Incident Update
Post an update to an open incident. This is the node that moves an incident through its lifecycle the way customers expect: the status you set here becomes the incident's status AND appears as a dated entry on the public timeline explaining the change. Posting RESOLVED closes the incident. The update is on the public page as soon as this returns, because an update always belongs to an incident that is already there, and turning notify on emails and messages every confirmed subscriber of the affected components. The statuses list can move component colours at the same time, which is how you take a component back to OPERATIONAL as part of resolving.
Instatus: Add Incident Update From Template
Post an incident update from a saved template, most often the resolution message. The template supplies the wording, the status and the component statuses, so an automation that watches a monitor recover can close the incident with the exact words your team agreed on. The template's own notify setting decides whether subscribers hear about it. This is the only operation in the incident-update family on /v2; its siblings are all on /v1, which is Instatus's split rather than a mistake here.
Instatus: Add Maintenance Update
Post an update on a maintenance window, which is how the window moves from NOTSTARTEDYET to INPROGRESS to COMPLETED with an explanation attached. Use it to say the work has started, that it is running long, or that it finished early. The entry is on the public page immediately, and notify emails and messages every confirmed subscriber of the affected components. If the window has autoStart or autoEnd on, Instatus will also move it by itself, so pick one mechanism rather than both.
Instatus: Add Metric Data Point
Push one data point onto a metric chart. The timestamp is epoch MILLISECONDS, so 1608672985000 rather than 1608672985, and a value in seconds lands in 1970 and quietly ruins the chart's axis. The point appears on the public chart immediately. Note the path: a POST to the metric itself adds a point, while a POST to the metric's /data adds many, so the two write endpoints differ only by that segment.
Instatus: Add Metric Data Points In Bulk
Push many data points onto a metric chart in one call, which is the node for backfilling history or for a scheduled job that reports a batch of measurements at a time. Every timestamp is epoch MILLISECONDS. The response echoes the accepted points, so compare its length against what you sent rather than assuming the whole batch landed.
Instatus: Add Monitors To Group
Put monitors into a group. Instatus documents this as adding the listed monitors and returns only a success message, so confirm the outcome with Instatus: List Monitors rather than assuming whether the list added to the group or replaced its membership; the response says nothing either way.
Instatus: Add Subscriber
Subscribe an email address or phone number to a status page, either to everything or to named components only. What happens next is decided by the page's subscriberConfirmationMode: REQUIRED sends a confirmation request they must accept, WELCOME adds them at once and sends a welcome email, NONE adds them silently, and the last two are Pro plan only. Setting autoConfirm on this call overrides the page setting for this one subscriber and behaves like NONE. Either way a real message may reach a real person, so do not point this at a list you have no consent for. Marked Destructive for endpoint testing on exactly that basis.
Instatus: Add Subscribers In Bulk
Subscribe up to 100 people to a status page in one call. Each entry takes the same fields as the single-subscriber node, so a batch can mix whole-page and per-component subscriptions and mix email with SMS. Partial success is normal and is reported honestly: the response carries created and failed counts plus a results object listing each created subscriber and each failure with its index, the entry that failed and an error code such as creation_failed for a duplicate. Check the failed array rather than trusting the HTTP status. autoConfirm applies to the whole batch and is Pro plan only.
Instatus: Add Teammate
Give someone access to a status page by email address. This sends them an invitation, so it reaches a real inbox the moment it runs, and it grants edit rights over a public page: the person can post incidents customers will see. Marked Destructive for endpoint testing because it mails a real person, not because it deletes anything.
Instatus: Create Audience Group
Create an audience group on a private status page, pairing a set of teammates with the components they are allowed to see. This is an access-control write: it decides which customers see which services on a shared private page. Note the underscore in the path. Instatus creates and updates audience groups under audience_groups while listing, reading and deleting them under audience-groups, and this catalog carries both spellings verbatim because either substitution returns 404. The response is a thin record without the teammates and components expanded; read the group back if you need those.
Instatus: Create Component
Add a component to a status page. It appears on the public page as soon as this returns, so a component created with a status other than OPERATIONAL announces a problem the moment it is created. To nest it under an existing group, set grouped to true and group to the group's id. Reads are on /v2 and this create is on /v1, which is Instatus's own split.
Instatus: Create Cron Monitor
Create a cron monitor for a scheduled job. period is how often the job should report in and grace is how much late is still acceptable, BOTH IN SECONDS, so a daily job with an hour of slack is 86400 and 3600. The response carries a slug, which is what your job pings; the background check starts on the first successful ping, so a monitor created and never pinged does not immediately alarm. As with uptime monitors, the onFail object decides whether a missed run raises a PUBLIC incident and notifies subscribers.
Instatus: Create Escalation Policy
Create an escalation ladder for unacknowledged incidents. Each rule waits delayInMins, checks its condition, then runs its actions, and the rules fire in order. The only condition Instatus documents is NOT_ACKNOWLEDGED. Two policy-level settings deserve a decision rather than a default: revertAcknowledgement takes an incident back to unacknowledged so the ladder can run again, and autoResolveIncidentsAfterRepeat closes the incident once the ladder has run out of people to page, which resolves an incident nobody actually fixed. Creating a policy pages nobody; the next unacknowledged incident does.
Instatus: Create General Notice
Add a standing banner to a status page for something that is neither an incident nor a maintenance: an upcoming migration, a deprecation, changed support hours. The banner is visible to visitors as soon as this returns. Unlike incidents, notify here defaults to false in Instatus's own documentation as well as in this node, so a notice is quiet unless you deliberately announce it.
Instatus: Create Incident
Open an incident on a status page. Read the two flags before wiring this into anything automatic. shouldPublish defaults to TRUE, which is Instatus's own default, so unless you set it to false the incident is on the public page the moment this returns. notify sends email, SMS and chat notifications to every confirmed subscriber of the affected components, and it starts off here because Instatus documents no default for it. Together those two decide whether this call is a private draft or a customer announcement. The status field sets the opening state of the lifecycle, normally INVESTIGATING; every later transition should be posted as an incident update so the public page gets a timeline entry explaining it. Naming components attaches them; the statuses list is what actually changes their colour. Marked Destructive for endpoint testing because it can mail a real subscriber list, not because it deletes anything.
Instatus: Create Incident From Template
Open an incident from a saved template, taking its name, message, opening status, affected components and their statuses from the template rather than from this node's config. This is the node for a rehearsed response: the wording was agreed in advance and the automation only decides when to post it. The template must be of type INCIDENT. Everything the template carries applies, including its own notify setting, so check the template before wiring this to a trigger. Returns the raw incident record rather than the hydrated shape the other incident nodes return.
Instatus: Create Metric
Create a metric chart on a status page, such as response time or requests per minute. The chart appears publicly as soon as this returns and stays empty until data points are pushed to it, so create it and feed it in the same workflow rather than leaving a blank chart on a customer-facing page.
Instatus: Create Monitor
Create an uptime monitor that checks a URL on a schedule from chosen regions and asserts on the response. The two objects worth reading before wiring this to anything are onFail and onRecover: they decide whether a failing check CREATES A PUBLIC INCIDENT, publishes it and notifies subscribers, and whether a recovery resolves it again. A monitor created with onFail.publishIncident and onFail.notifySubscribers on is a standing instruction to tell your customers about an outage without a human in the loop, which is usually the point and is occasionally a very expensive mistake. createComponent and createMetric add a component and a response-time chart to the public page alongside the monitor. Note that pageId goes in the BODY here, because the create path carries no page segment.
Instatus: Create Monitor Alert
Create an alert that delivers monitor state changes to a person or channel. The type decides the delivery route and the recipient is read in terms of it: an email address for EMAIL, a phone number for SMS or PHONE_CALL, a channel or webhook target for the chat types. INCIDENT is different from all of them, because it raises an incident on the status page rather than messaging anybody directly. Creating the alert sends nothing; the first matching monitor state change does. whenSslExpires with sslExpiresInDays is the certificate warning, which is the one alert here that fires on a schedule rather than on a failure.
Instatus: Create Monitor Group
Create a group to organise monitors, optionally nested under another group. Groups collapse a long monitor list into something readable and can be paired with a component on the public page. Note the path spelling: monitors-groups, plural on both words. Instatus ships no endpoint that lists groups, so record the id this returns; nothing else will give it back to you through the API.
Instatus: Create Monitoring Integration
Connect a third-party monitoring tool such as Pingdom or New Relic to a status page, so alerts from that tool land on named components. This is the family that arms automatic incident creation from outside Instatus: once it exists, an alert from the connected tool can put an incident on the public page with nobody in the loop. These are the only three operations on /v3 in the entire API, which is Instatus's own versioning and not a typo here. Instatus documents no endpoint that LISTS integrations, so record the id this returns.
Instatus: Create On-call Schedule
Create an on-call schedule with one or more shifts, each holding its own members, rotation and time restrictions. This is the richest body in the connector and it is nested three deep: a schedule holds shifts, a shift holds members and time intervals. Rotation is DAILY, WEEKLY or CUSTOM, and a CUSTOM rotation needs customRotationTypeValue and customRotationTypeUnit. Restrictions are NONE for round the clock, TIMEOFDAY for the same hours every day, or TIMEINTERVALS for named day and time spans such as Saturday 21:00 to Sunday 03:00. Record the id this returns: there is no endpoint that lists schedules.
Instatus: Create Routing Rule
Create a rule that decides who gets notified about an incoming event. Assertions match on the event, actions say who to notify, and order decides which rule wins. Note two shapes specific to this family. The whole body is WRAPPED in a routingRule object, which this node does for you. And an action's targetId means different things per action type: a workspace member EMAIL ADDRESS for NOTIFY_WORKSPACE_MEMBER, a team id for NOTIFY_TEAM, and nothing at all for the on-call types, which resolve their target from the schedule. Creating a rule notifies nobody; the next matching event does.
Instatus: Create Status Page
Create a new status page with its first components in one call. The page goes live at subdomain.instatus.com as soon as this returns, so pick the subdomain deliberately: it is the public address customers will see. Passing a list of component names creates those components alongside the page, which is the quickest way to stand up a working page. Note the version split here is real and not a typo: pages are LISTED on /v2 and CREATED on /v1.
Instatus: Create Template
Save a reusable incident or maintenance template so a response can be posted later in wording your team agreed in advance. Creating a template publishes nothing and notifies nobody; the notify flag stored here takes effect when the template is POSTED. The components field on this endpoint is a list of OBJECTS carrying an id, a status and optionally a customImpactPercentage, which is different from the plain list of id strings the incident and maintenance endpoints take.
Instatus: Create Workspace
Create a workspace to hold status pages and members, and set the notification defaults every new member inherits. Useful for agencies and platform teams standing up a separate container per client or per product line.
Instatus: Delete Audience Group
Permanently delete an audience group. There is no trash and no undo, and everyone in the group loses the scoped view it gave them. Back to the HYPHEN spelling here: delete uses audience-groups while create and update use audience_groups.
Instatus: Delete Component
Permanently delete a component from a status page, along with its recorded uptime history. There is no trash and no undo, and any subscriber who subscribed to that component specifically loses that subscription. If the goal is to stop showing a component while keeping its history, update it with archived set to true instead. Reads are on /v2 and this delete is on /v1.
Instatus: Delete Cron Monitor
Permanently delete a cron monitor and its ping history. There is no trash and no undo, and the ping slug stops working, so the job carries on pinging into nothing. Nobody is told the job is now unwatched. If the goal is to quieten it temporarily, update its state to PAUSED or MUTED instead.
Instatus: Delete Escalation Policy
Permanently delete an escalation policy and its whole ladder. There is no trash and no undo. Incidents that would have escalated through it now do not escalate at all, which is a failure you discover during the next outage rather than now.
Instatus: Delete General Notice
Permanently delete a general notice, removing the banner from the public page. There is no trash and no undo. This is the normal way to take a notice down once the thing it announced is over.
Instatus: Delete Incident
Permanently delete an incident and every update on it. There is no trash and no undo, and the incident disappears from the public history as though it never happened, which changes the page's uptime record. Notifications already sent to subscribers are not recalled. Deleting is for incidents raised in error; a real incident that is over should be RESOLVED rather than removed.
Instatus: Delete Incident Update
Permanently remove one entry from an incident's public timeline. There is no trash and no undo. Deleting the most recent update does not roll the incident's status back to what the previous entry said, so an incident can be left claiming a status nothing on its timeline explains. Notifications already sent for that update are not recalled.
Instatus: Delete Maintenance
Permanently delete a maintenance window and every update on it. There is no trash and no undo. If subscribers already received the advance warning, deleting the window does not tell them it is cancelled: post a maintenance update saying so first, then delete, or the calendar entry they made stands.
Instatus: Delete Maintenance Update
Permanently remove one entry from a maintenance window's public timeline. There is no trash and no undo, and removing the most recent entry does not roll the window's status back, so it can be left claiming a state nothing on its timeline explains.
Instatus: Delete Metric
Permanently delete a metric and every data point recorded against it. There is no trash and no undo, and the chart disappears from the public page. To clear the numbers while keeping the chart, use Instatus: Delete Metric Data Points instead.
Instatus: Delete Metric Data Points
Permanently delete the data points on a metric while keeping the metric itself. There is no trash, no undo and no way to delete a single point or a range: this clears the series. Use it when a bad feed has poisoned a public chart and the fastest honest fix is to empty it and re-push.
Instatus: Delete Monitor
Permanently delete a monitor and its check history. There is no trash and no undo. Checking stops at once, which means the thing it watched is now unwatched and nothing will say so: a deleted monitor produces silence, not an alert. Any component or metric it created on the public page is left behind.
Instatus: Delete Monitor Alert
Permanently delete a monitor alert. There is no trash and no undo. The monitors it covered keep running and keep recording failures; they simply stop telling anyone through this route, so the symptom of a wrong delete here is silence during the next outage rather than an error now.
Instatus: Delete Monitor Group
Permanently delete a monitor group. There is no trash and no undo. Instatus does not document what becomes of the monitors inside it, so treat their fate as unknown and confirm they are still running afterwards with Instatus: List Monitors rather than assuming they were simply ungrouped.
Instatus: Delete Monitoring Integration
Permanently disconnect a third-party monitoring tool from a status page. There is no trash and no undo. The external tool keeps monitoring and keeps alerting its own way; what stops is those alerts reaching this status page, so the visible symptom is a page that no longer reflects an outage the tool can see. Returns an empty object on success.
Instatus: Delete On-call Schedule
Permanently delete an on-call schedule and every shift on it. There is no trash and no undo. Any routing rule or escalation policy action of type NOTIFY_CURRENT_ON_CALL_USER, NOTIFY_NEXT_ON_CALL_USER or NOTIFY_ALL_ON_CALL_MEMBERS resolves its target from a schedule, so deleting one can leave those actions with nobody to page, and they will not say so until an incident needs them. Returns HTTP 204 with an empty body, so a successful run shows data as an empty object.
Instatus: Delete Outage
Permanently delete a recorded outage period. There is no trash and no undo, and the component's published uptime percentage goes UP as a result, because the downtime it was counting stops existing. That makes this the one delete in the connector that improves a public number, which is exactly why it deserves a deliberate hand on it. Returns HTTP 204 with an empty body, so a successful run shows data as an empty object.
Instatus: Delete Routing Rule
Permanently delete a routing rule. There is no trash and no undo. Events that matched it now fall through to whatever rule comes next, or to nothing, so the consequence of a wrong delete here is an unattended incident rather than an error message.
Instatus: Delete Status Page
Permanently delete a status page. This takes the public page offline and removes its components, incidents, maintenances, metrics, subscribers and recorded uptime history with it. There is no trash and no undo, and the subdomain stops resolving immediately for anyone who has the link. Same path shape as the update: /v2/{page id}, no /pages segment.
Instatus: Delete Template
Permanently delete a saved template. There is no trash and no undo. Nothing already posted from it changes, but any automation that references this template id starts failing, which is the failure mode worth checking for first: the template is a dependency of every workflow that posts it.
Instatus: Delete Workspace
Permanently delete a workspace. This is the widest-blast-radius call in the connector: a workspace contains status pages, and those pages carry their own incidents, subscribers and history. There is no trash and no undo. Confirm the workspace has no pages you still need before running it.
Instatus: Edit Incident Update
Edit an update already on the public timeline: fix its wording, its status, its timestamp or the component statuses it applied. Customers who already read the original or received it by email see no correction unless you turn notify on, so an edit is not a retraction. For a material change of story, post a new update rather than rewriting history.
Instatus: Edit Maintenance Update
Edit an update already on a maintenance window's public timeline: its wording, status, effective time or the component statuses it applied. Subscribers who already received the original are not told about the correction unless notify is turned on.
Instatus: Ping Cron Monitor Failure
Tell Instatus a scheduled job failed, explicitly and immediately, rather than waiting for its window to lapse. Put this on the error branch of an automation so a failure is reported the moment it happens instead of up to a period plus a grace later. This applies the cron monitor's own onFail behaviour, so on a monitor configured to publish and notify it raises a public incident and messages subscribers. Marked Destructive for endpoint testing on that basis.
Instatus: Ping Cron Monitor Start
Tell Instatus a scheduled job has started. Pair it with a success or failure ping at the end and Instatus records how long the run took, which turns a cron monitor from did it run into how long did it take. Sending a start ping with no matching finish leaves the run open, so put the finishing ping on every branch that ends the job, including the error branch.
Instatus: Ping Cron Monitor Success
Tell Instatus a scheduled job finished successfully. Put this at the end of any Mattering automation that has to run on time, and Instatus will notice when it stops arriving. The slug is the secret: it comes back from Instatus: Create Cron Monitor and from Instatus: List Cron Monitors, and anyone who has it can report on that job's behalf, so treat it as a credential rather than as an id. The first successful ping is what starts the background check, so a new cron monitor is not armed until this runs once.
Instatus: Record Outage
Record a downtime period against a component directly, without an incident. This is the backfill node: it is how you enter history from before you used Instatus, or correct a period the monitors missed. Recording an outage CHANGES THE PUBLISHED UPTIME PERCENTAGE for that component, so it is publicly visible in the numbers even when visibilityState keeps the period itself out of the timeline. It sends no notifications.
Instatus: Regenerate Secure Link
Mint a new secure link for a private status page and invalidate the old one. Everyone holding the previous link loses access the moment this returns, which is the point when a link has leaked and also the reason not to run it casually. The page must be private with secure links enabled, and your account needs member or owner access to it. The link embeds a JWT, so treat the returned value as a credential.
Instatus: Remove Subscriber
Unsubscribe somebody from a status page. They stop receiving notifications immediately and there is no undo: resubscribing them means a new confirmation cycle. This is the node to use for an unsubscribe or deletion request, and it is also the node that quietly cuts a customer off from outage notices if pointed at the wrong id.
Instatus: Remove Teammate
Remove somebody's access to a status page. Their access ends immediately. Incidents and updates they already posted stay on the page and keep their name attached. Pass the TEAMMATE id from the list endpoint, which is the outer id, not the nested user id.
Instatus: Run Monitor Check
Run one monitor's check immediately instead of waiting for its schedule, optionally from a named region. This is a GET that DOES something, which is why it sits with the actions. It sends a real request to the monitored target, and if the check fails the monitor's own onFail behaviour applies, so a monitor configured to publish and notify can raise a public incident as a result of this call. Use it to confirm a fix has landed before the next scheduled check.
Instatus: Run Monitor Group Check
Run every monitor in a group immediately, optionally from a named region. Another GET that does something, which is why it sits with the actions. Each monitor's own onFail behaviour applies, so this can raise several public incidents at once if the group is configured that way and the target is genuinely down. Useful after a deploy to confirm a whole surface is healthy without waiting for the schedule.
Instatus: Schedule Maintenance
Schedule a maintenance window on a status page. It is visible to visitors as soon as this returns, and it is expanded on the page by default; set isCollapsed to true with an expandAt time to have it appear quietly and open itself closer to the window. Four separate notification switches decide who hears what and when: notify announces the schedule now, notifyEarly with notifyMinutes warns subscribers that many minutes before it starts, notifyStart tells them it has begun and notifyEnd tells them it is over. autoStart and autoEnd move the window from NOTSTARTEDYET to INPROGRESS to COMPLETED on the schedule with no further API calls, so an automation that also posts those transitions will produce them twice unless the flags are off. Marked Destructive for endpoint testing because it can mail a real subscriber list.
Instatus: Update Audience Group
Update an audience group's name, components or teammates. A list sent here REPLACES the existing one rather than adding to it, so read the group first and send the full intended set, or you will silently revoke access for everyone you left out. Underscore in the path, as on create.
Instatus: Update Component
Update a component's name, description, status, order, grouping or archived flag. Changing status is publicly visible immediately and recalculates the page's overall status, so setting MAJOROUTAGE here turns the page red without creating an incident or telling anyone why. For a customer-facing outage, prefer creating an incident with a statuses entry, which sets the component status AND leaves a timeline explaining it. Note the group field is spelled group on create and groupId here.
Instatus: Update Cron Monitor
Update a cron monitor's name, schedule, grace, alerts or behaviour, or change its state. The state field is the useful one for automation: PAUSED stops checking entirely and MUTED keeps checking while sending no notifications, which is how you silence a known-broken job during a planned migration without losing its history. Remember that period and grace are seconds.
Instatus: Update Escalation Policy
Update an escalation policy's name, ladder or repeat settings. The rules list replaces the existing ladder rather than merging with it, so send the whole thing: a short list silently removes the later rungs, and the way you find out is that nobody was paged after the first one did not answer.
Instatus: Update General Notice
Update a general notice. Instatus applies only the fields you send, so this is a genuine partial update and omitting a field leaves it alone. The edit is live on the public page immediately.
Instatus: Update Incident
Edit an incident record: its title, start time, affected components and their statuses. Understand what this does NOT do before using it to move the lifecycle along. Changing status here edits the record without adding a timeline entry, so a page can end up reading RESOLVED with no message saying how it was resolved. To move an incident from INVESTIGATING to IDENTIFIED to MONITORING to RESOLVED the way customers expect, post an incident update instead. Use this node for corrections: a wrong title, a wrong start time, a component that should not have been attached.
Instatus: Update Maintenance
Reschedule or edit a maintenance window: its title, times, duration, notification flags and covered components. As with incidents, changing status here edits the record without adding a timeline entry, so use a maintenance update when you want the public page to say WHY the window started late or ran long. Moving the start time of a window whose advance warning has already gone out does not un-send it.
Instatus: Update Metric
Rename a metric or change the unit suffix shown after its values. The change is live on the public chart immediately. Changing the suffix does not convert the stored numbers, so relabelling milliseconds as seconds makes the chart wrong rather than converted.
Instatus: Update Monitor
Update a monitor: its target, method, headers, assertions, check interval, region, alerts or its onFail and onRecover behaviour. Turning onFail.publishIncident or onFail.notifySubscribers ON here is a customer-facing change even though nothing is published by this call: the next failed check is what acts on it, unattended.
Instatus: Update Monitor Alert
Update a monitor alert's channel, recipient, trigger conditions or the monitors it covers. The monitors list REPLACES the existing one, so sending a short list silently stops alerting for every monitor left out, which is a failure that announces itself only by nothing happening during the next outage. Read the alert first.
Instatus: Update Monitor Group
Rename a monitor group, collapse it, reorder it or pair it with a component on the public status page. Nothing here is visible to customers except the component pairing, which decides which public component the group's monitors report against.
Instatus: Update Monitoring Integration
Update a monitoring integration: any field the create endpoint takes, plus a display label and the two template objects that customise what an incoming alert actually says. createTemplate shapes the incident this integration opens and resolveTemplate shapes how it closes, which is how you stop a third-party tool posting raw alert text on a customer-facing page.
Instatus: Update On-call Schedule
Update an on-call schedule's name, description, status or shifts. Instatus's own example sends the ENTIRE schedule back, ids and all, including every shift, member and time interval, so the safe pattern is to read the schedule, change what you need and send the whole object rather than a partial one. Dropping a shift or a member from that object takes them off the rotation, which means nobody is paged for the window they used to cover.
Instatus: Update Outage
Correct a recorded outage period: its severity, its start, its end or its visibility. Closing an open period by setting to is the common case. Every one of these edits moves the published uptime percentage for the component, so treat it as a correction to a public number rather than as an internal edit.
Instatus: Update Routing Rule
Update a routing rule's assertions, actions or evaluation order. As with create, the body is wrapped in a routingRule object. The lists replace rather than merge, so a short actions list quietly stops paging everyone left out, and the symptom is nobody being woken up during the next incident. Read the rule first with Instatus: List Routing Rules.
Instatus: Update Status Page
Update a status page's name, subdomain, branding, custom HTML and CSS, privacy settings and subscriber confirmation behaviour. Note the path: this operation is /v2/{page id} with no /pages segment at all, which is Instatus's own spelling and not an omission here. Changing the subdomain changes the public address customers already have, so treat it as a migration rather than an edit.
Instatus: Update Template
Edit a saved template's type, name, message, posted status, notify flag or component statuses. Nothing is published and nobody is notified by this call. Changing the notify flag changes the behaviour of every automation that posts this template afterwards, which is the one edit here with a customer-facing consequence.
Connecting Instatus takes one credential
Instatus issues API keys from the dashboard, under User settings and then Developer settings at dashboard.instatus.com/developer. There is no OAuth flow and no per-scope selection: a key carries whatever the user carries, which is why the choice of which Instatus user creates it matters more than it looks. A key created by a person with owner access to every status page in a workspace can delete any of them. A key created by a teammate with access to one page can only touch that page. The API is available on every Instatus plan including the free Starter tier, so a key costs nothing to obtain.
Decide which Instatus user holds the key. Every node runs as that person, and Instatus: Get User Profile will tell you at any time whose key a workflow is using. An account the team controls is better than an individual's login for anything unattended, because a departing employee should not silently break your incident automation.
Sign in to the Instatus dashboard, open User settings, then Developer settings at dashboard.instatus.com/developer, and create an API key. Copy it once and store it in Mattering rather than in a spreadsheet.
In Mattering Automations, open Connectors, click Add Connector and choose Instatus, then click Create new credential and paste the key. It is stored encrypted and never appears in workflow logs.
Click Test Connection. Mattering asks Instatus for the first page of your status pages, which is the cheapest read on the API and takes no parameters. If the key is wrong you get an HTTP 401. Read the status rather than the message body: Instatus answers a rejected key with a 401 whose body says forbidden, which is 403 vocabulary, and the status line is the one telling the truth.
All 109 Instatus nodes appear in the canvas palette. Run Instatus: List Status Pages first. It returns the ids every other node wants and populates every page picker in the connector, so a working list here means the pickers work everywhere else. An empty array is a valid answer rather than a failure: Instatus keys belong to a person, and a person with no pages sees none.
Before you build anything customer-facing, consider making a throwaway status page on the free plan with no subscribers on it. Instatus: Create Status Page will do it in one call. It gives you somewhere to run the incident and maintenance nodes for real without an audience watching.
Common uses for the Instatus MCP
Turn an alert from anywhere into a published incident
Most teams already detect outages somewhere that is not their status page. A webhook from an alerting tool, a failed health check in a workflow, or a support queue spiking can all run Instatus: Create Incident with the affected components named and the opening status set to INVESTIGATING. Read the two flags first: shouldPublish defaults to true, so the incident is on the public page the moment the node returns, and notify decides whether every confirmed subscriber of those components gets an email, an SMS or a chat message. Instatus: Create Incident From Template is the version worth reaching for once the wording has been agreed in advance, because then the automation only decides when to post, not what to say.
Watch a supplier's status page without any relationship with the supplier
Instatus: Get Public Page Summary is the only node here that needs no credential, because it reads the summary any Instatus status page publishes at its own address. Point it at a vendor you depend on, run it on a schedule, and you can react to their outage before your support queue tells you about it: open a Mattering Productivity task, post to a channel, or set your own component to DEGRADEDPERFORMANCE with Instatus: Update Component. It returns the page's overall status plus its active incidents and active maintenances, using the condensed vocabulary of UP, HASISSUES and UNDERMAINTENANCE. This works on any Instatus page, and it works whether or not you have an Instatus account of your own.
Make every scheduled job prove it ran
A cron monitor catches the failure mode alerting usually misses: a job that stops running entirely, which produces no error to alert on. Create one with Instatus: Create Cron Monitor, noting that its period and grace are both in seconds, then put Instatus: Ping Cron Monitor Success at the end of the automation that has to run on time. If the ping stops arriving, Instatus notices. Instatus: Ping Cron Monitor Start at the beginning turns the question from did it run into how long did it take, and Instatus: Ping Cron Monitor Failure on the error branch reports a failure immediately instead of waiting for the window to lapse. The ping slug is effectively a credential, since anyone holding it can report on that job's behalf.
Announce planned maintenance from the system that scheduled it
A deploy pipeline, a change calendar or an approved Productivity task can call Instatus: Schedule Maintenance so the window appears on the public page with the right components attached and the advance warning already set. Four separate notification switches decide who hears what and when, and autoStart and autoEnd will move the window from NOTSTARTEDYET to INPROGRESS to COMPLETED on the schedule with no further calls. Pick one mechanism: an automation that also posts Instatus: Add Maintenance Update transitions on a window with autoStart on will produce each transition twice. Use the update node for the things a schedule cannot know, such as the work running long or finishing early.
Publish numbers your customers can check
Instatus: Create Metric adds a chart to the public page and Instatus: Add Metric Data Point feeds it one measurement at a time, with Instatus: Add Metric Data Points In Bulk for backfilling history or reporting a batch on a schedule. Timestamps are epoch milliseconds rather than seconds, which is the one detail that quietly ruins a chart's axis when it is wrong. A workflow that reads a latency figure from your own monitoring and pushes it here every few minutes turns a status page from a place people go during an outage into a page that says something on a normal day. If a bad feed has poisoned a chart, Instatus: Delete Metric Data Points clears the series without deleting the chart.
Audit who gets paged and who gets told
The on-call and subscriber sides of Instatus are usually configured once and then trusted for a year, which is how a rotation ends up pointing at somebody who left. All of it is readable. Instatus: List On-call Schedule Members answers who is on call right now and who is next, which makes it worth wiring into a Matter agent so anyone can ask. Instatus: List Routing Rules and Instatus: List Escalation Policies show who gets notified for what and in which order. Instatus: List Teammates shows who can post to the public page, and Instatus: List Subscribers shows who receives the notifications. That last payload is a customer contact list, so treat what a workflow does with it accordingly.
Example: A supplier outage that updates your own status page with a human in the loop
Built on the Automations canvas in a few minutes, or described in plain language to Matter and built for you.
Trigger: a schedule fires every few minutes. This connector has no trigger nodes, so Instatus workflows run on a schedule or off another system's event rather than listening for a push from Instatus.
Instatus: Get Public Page Summary reads a supplier's status page at its own address. No credential is involved in this step, because the summary is public.
A condition checks whether the returned status is still UP. If it is, the run ends here and nothing has been published anywhere.
If it is not, Instatus: Update Component sets your own dependent component to DEGRADEDPERFORMANCE. This is publicly visible immediately and it changes the page colour, but it creates no incident and notifies nobody, which is the right amount of signal while a person decides what to say.
A task is created in Mattering Productivity with the supplier's status, their active incidents and a draft customer message, so somebody who understands the blast radius reads it before your subscribers do.
On approval, Instatus: Create Incident From Template posts the agreed wording. The template carries the status, the affected components and its own notify setting, so the decision about whether to mail the subscriber list was made when the template was written rather than in a hurry during the outage.
When the supplier recovers, Instatus: Add Incident Update From Template posts the resolution, which sets the incident to RESOLVED and puts a dated entry on the public timeline explaining it.
Instatus MCP questions, answered
What can this connector actually do in Instatus?
109 nodes: 33 that read and 76 that act. Reads cover status pages, workspaces, components, incidents and their updates, maintenance windows and their updates, templates, general notices, recorded outages, teammates, subscribers, metrics, audience groups, the user profile behind the key, the public page summary, uptime monitors and their check logs, monitor alerts, cron monitors with their logs and summary, routing rules, escalation policies and on-call schedule members. Writes cover creating, updating and deleting all of the above, plus the three cron ping endpoints, running a monitor check on demand, regenerating a private page's secure link, and adding subscribers singly or up to 100 at a time. Two things are worth knowing before you plan. This connector ships no trigger nodes, so Instatus workflows run on a schedule or off another system's event rather than being pushed from Instatus. Instatus can deliver notifications to a webhook subscriber, so a status page can be pointed at a webhook trigger in Mattering, but that is a subscription you configure on the Instatus side rather than a node in this list. And a few families have no lister on the Instatus side at all, specifically monitor groups, monitoring integrations and on-call schedules, so their ids are free text rather than pickers and come from the create response or the dashboard URL.
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 work with Instatus through them. Matter, the assistant built into the platform, uses the same nodes natively. The same nodes are also steps on the automation canvas, so a workflow an agent helps you design can then run on a schedule with nobody present. Given that this connector writes to a public page, the sensible split is to hand an assistant the 33 read nodes freely, keep the 38 flagged destructive behind a condition or an approval step, and be especially deliberate about the ones that mail subscribers. An agent that can answer who is on call and what is currently broken is useful with no risk attached; an agent that can post an incident is a different decision.
Does Instatus publish its own MCP server?
No. There is no first-party Instatus MCP server. Instatus's help site and API reference describe the REST API and do not mention MCP or the Model Context Protocol anywhere, and the addresses a vendor server would normally live at return 404. So this connector is the practical way to put Instatus behind an MCP-capable assistant today, and it brings things a local server config would not: the nodes run on a schedule or off another system's event with nobody present, they chain with every other connector in one workflow so an alert from your monitoring can become an incident here, the credential lives once in a shared workspace store rather than in each person's machine, and the 38 nodes that publish, mail or delete are flagged as such before anyone runs them.
Will an automation post something my customers can see?
Yes, and this is the question to answer before you build anything. Instatus: Create Incident publishes to the live status page by default, because shouldPublish defaults to true in Instatus's own documentation and this connector preserves the vendor's default rather than quietly inverting it. The notify flag, wherever it appears on incidents, incident updates, maintenances, maintenance updates and templates, sends email, SMS and chat notifications to every confirmed subscriber of the affected components. Instatus documents no default for notify anywhere, so the connector sends it explicitly on every request and starts it off, because an undocumented default that mails a customer base is not something to discover in production. General notices are the exception in the other direction: Instatus documents notify as defaulting to false there, and that node keeps the vendor's default. Beyond those flags, Instatus: Update Component changes the page's colour immediately with no incident and no notification, Instatus: Create Metric puts an empty chart on the page, and a monitor created with its onFail behaviour set to publish and notify is a standing instruction to tell your customers about an outage with no human in the loop. The safest way to learn any of this is a throwaway page on the free plan with no subscribers on it.
Do I need a paid Instatus plan?
No. Instatus's pricing page says of API access, in its own words, that it is available on all plans, and the Starter plan is free forever with no card required, so you can create a key and run this connector without paying anything. Three limits are worth knowing because they affect specific nodes rather than the API as a whole. The WELCOME and NONE subscriber confirmation modes are documented as Pro only, and so is the autoConfirm override on Instatus: Add Subscriber and Instatus: Add Subscribers In Bulk, so on a free plan a new subscriber always gets a confirmation request. Private and select-audience status pages are an upgrade beyond Starter, whose status page type is listed as Public, which means the five audience-group nodes and Instatus: Regenerate Secure Link need an upgraded page to act on. And the Starter plan's own caps apply to what the nodes can create: 15 monitors, 200 subscribers, 5 team members and 2 on-call members. Nothing else observed is plan-gated, and no node returns an empty list to disguise a plan boundary.
What are Instatus's rate limits?
Instatus does not document a rate limit in its API reference, so we are not going to invent a number for it. What the connector does instead is treat an HTTP 429 as what it is: a throttle rather than a credential failure, reported as such rather than as a broken connection. Two practical habits matter more than a published figure here. Each node is one request and the connector does not auto-paginate, so an unbounded walk never hides behind a single canvas step and you decide when to ask for the next page. And the versioned half of this API returns a bare JSON array with no total, so there is no way to tell from one response whether more rows exist: you ask for the next page and stop when one comes back empty. The monitoring half is different and returns total, page, totalPages and limit alongside the rows, which is worth knowing when you are writing the loop.
Can I monitor a status page I do not own?
Yes, and it is one of the more useful things here. Instatus: Get Public Page Summary reads the summary that every Instatus status page publishes at its own address, which is public data and needs no credential and no relationship with the page's owner. Give the node a status page URL and it returns that page's overall status plus its active incidents and active maintenances. So you can watch the vendors your product depends on and react to their outage on your own page or in your own tools before your support queue tells you about it. Two details: the status vocabulary on this endpoint is the condensed three-value list of UP, HASISSUES and UNDERMAINTENANCE rather than the longer list the authenticated API returns, and this only works for status pages hosted by Instatus, since it is Instatus's own published format rather than a general web scraper.
How should an automation move an incident through its lifecycle?
Post an update, not an edit. This is the single most common way to end up with a status page that reads wrong. An incident's lifecycle runs INVESTIGATING to IDENTIFIED to MONITORING to RESOLVED, and Instatus: Create Incident sets only the opening status. Every transition after that should be Instatus: Add Incident Update, because that is what produces a dated entry on the public timeline explaining the change. Instatus: Update Incident will also change the status field, but it adds no timeline entry, which is how a page ends up saying RESOLVED with nothing on it explaining how. Use that node for corrections instead: a wrong title, a wrong start time, a component that should not have been attached. Maintenance windows work the same way, with the added wrinkle that autoStart and autoEnd move a window through NOTSTARTEDYET, INPROGRESS and COMPLETED by themselves, so an automation posting those transitions as well will double them up. And note that deleting the most recent update does not roll the parent's status back to what the previous entry said, which leaves an incident claiming a status nothing on its timeline explains.
Put Instatus 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. Instatus is a product of its respective owner; Mattering is an independent platform.