Instatus MCP. Manage Incidents and Components from Chat.

Listen up, pal. Your agent connects to Instatus so you don't gotta touch a dashboard when something breaks. You manage your entire service status page—from reporting an outage to tracking down why the database is running slow—all through natural talk. It’s basically giving your AI client total control over observability data management.

Incident Management and Outages:
When things go sideways, you've got create_incident to report a brand new service outage right away. If you wanna make it look official, use create_incident_from_template. You can then manage the fallout by running update_incident, which lets your agent change the severity or mark an existing problem as resolved.

Need to clean up history? Use delete_incident to manually remove a recorded incident report. To see what’s going on, you pull everything with list_incidents, and if you need the deep dive on one specific issue, you run get_incident.

Component Status Control:
Every service—the API, the database, the website—is a component. You can define these new parts using create_component or see everything with list_components. When something changes, use update_component to change that specific component’s status—like flipping it from OPERATIONAL to DEGRADED. If you misconfigured something, delete_component gets rid of the definition entirely.

To check what a single service is doing right now, run get_component.

Scheduling and Maintenance Windows:
You gotta let people know when you're taking it down for updates. You can schedule future downtime using create_maintenance, which sets up a whole window on the page. If those plans change, use update_maintenance to shift the start or end date, or if you gotta bail, run delete_maintenance.

To see all scheduled work, you get a list with list_maintenances, and for details on one specific maintenance block, you check with get_maintenance.

Tracking Performance Metrics:
This is where you track the numbers. You can define entirely new performance metrics using add_metric. Once that metric exists, your agent records a single data point against it via add_metric_datapoint, or if you got a bunch of readings at once, use add_metric_datapoints_bulk to dump them all in one go.

To see what metrics you're tracking overall, run list_metrics. You can also clean up old data using delete_metric_datapoints. If you need the current operational status for a service or component, use get_component and list_components to pull that info.

User and Access Management:
You gotta control who sees this stuff. You can add new viewers by running add_subscriber, or if you're rolling out access to a bunch of people, use add_subscribers_bulk. To kick someone off the page, use remove_subscriber or list_subscribers to see who’s currently subscribed. For team roles, you can add new managers using add_teammate, or if someone loses their job, run remove_teammate.

You get a list of all authorized staff with list_teammates.

It's that straightforward. Your agent executes every single one of these actions without you ever having to open a browser tab.