Statuspage to MCP: Incidents, Maintenances, Components, Uptime and Subscribers in any AI Agent

💡 No install? Use cloud.anythingmcp.com directly. Sign in, click Connectors → Statuspage, paste your credentials, mint an MCP API key and you're done. No Docker, no git clone, no local server.

Statuspage

Atlassian Statuspage through its management API: status pages, components and groups with uptime, incidents and scheduled maintenances with every update, incident templates, subscribers and system metrics. Open and update incidents, announce maintenances, change component statuses, manage subscribers and push metric data.

Setup

Setting up takes about 2 minutes. You need a Statuspage account.

1. Sign in at https://manage.statuspage.io, open your avatar menu (bottom left) and choose API info.
2. Copy an existing API key or create one. The key acts with the rights of the user it belongs to and reaches every page of the organization; it cannot be limited to read-only.
3. Paste the key alone into STATUSPAGE_API_KEY, without OAuth or Bearer in front: the connector sends Authorization: OAuth <key> itself.
4. Install the connector and run statuspage_list_pages to check the key. Every other tool needs the page_id it returns; the page id is also shown on the API info page.

The connector uses the management API https://api.statuspage.io/v1, not the public status API of a page (<page>.statuspage.io/api/v2).

What is down right now: statuspage_list_unresolved_incidents (investigating, identified or monitoring), statuspage_list_active_maintenances, statuspage_list_components with each component's status, then statuspage_get_incident for every update. statuspage_list_incidents lists the history and searches names, statuses, postmortems and update texts with q. statuspage_get_postmortem reads the postmortem of an incident.

Components and uptime: statuspage_get_component, statuspage_create_component and statuspage_update_component, which changes a status without opening an incident. Groups with statuspage_list_component_groups, statuspage_get_component_group, statuspage_create_component_group and statuspage_update_component_group (it needs the name and the full component list on every update). statuspage_get_component_uptime and statuspage_get_component_group_uptime give the uptime percentage, major and partial outage seconds and the related events for a window of at most six calendar months (default the last 90 days); the component's uptime display must be enabled.

Incident communication: the AI can start from statuspage_list_incident_templates for agreed wording, open an incident with statuspage_create_incident (name, first message, status, affected components with their new status), post progress with statuspage_update_incident (identified, monitoring, then resolved, with a body each time and the components back to operational) and fix a typo in an earlier update with statuspage_edit_incident_update. New templates come from statuspage_create_incident_template.

Planned work: statuspage_create_scheduled_maintenance announces a window (scheduled_for to scheduled_until) with its components, optional reminders and automatic start and completion that move the components to under_maintenance and back. statuspage_list_upcoming_incidents and statuspage_list_scheduled_incidents show what is planned; statuspage_update_incident moves a maintenance to in_progress, verifying or completed, or reschedules it.

Subscribers: statuspage_count_subscribers and statuspage_get_subscriber_histogram for numbers per type and state, statuspage_list_subscribers (search with q), statuspage_get_subscriber, statuspage_list_unsubscribed_subscribers and statuspage_list_incident_subscribers. statuspage_create_subscriber subscribes someone by email, SMS or webhook, for the whole page or some components; statuspage_unsubscribe_subscriber removes them. Contact details are personal data, so the AI only fetches them when needed.

Metrics: statuspage_list_metrics, statuspage_get_metric and statuspage_list_metrics_providers (Datadog, New Relic, Pingdom, Librato or self-submitted). statuspage_add_metric_data pushes one value at a Unix timestamp to a metric whose data you submit yourself.

Before writing: incidents, maintenances and component statuses are public, and by default every change emails, texts and calls the webhooks of all subscribers. The AI shows you the wording and the components and waits for your yes; with deliver_notifications: false a change stays quiet.

Switched off at install: statuspage_delete_incident, which removes an incident or maintenance with all its updates from the page and its history and cannot be undone. Resolving the incident is usually the better choice; a workspace admin turns the tool on when the team needs it.

Data formats: ids are 12-character strings (for example gytm4qzbx9t6). Times are ISO 8601 with time zone; uptime dates are YYYY-MM-DD in the page's time zone and outage durations are seconds. Component status is operational, under_maintenance, degraded_performance, partial_outage or major_outage. Incident impact is none, maintenance, minor, major or critical, and incident_updates is newest first. Page colours and logo URLs are left out of the page answers.

Paging: lists take page (from 1) and per_page (at most 100); statuspage_list_incidents uses limit instead of per_page. statuspage_list_subscribers counts page from 0 and returns everything without limit, so the AI sets one on large pages. A page shorter than the page size is the last one.

Limits and common errors: each key may make about 1 request per second, measured over a minute; 420 or 429 means wait and retry. 401 is a wrong or deleted key, 403 a user without the rights or a plan without the feature, 404 an unknown id or a page the key cannot reach, 422 a validation error naming the field (for example a status that does not fit the incident type).

The connector does not change page settings, does not delete components, groups or templates, does not write or publish postmortems and does not create metrics or metric providers.

Prerequisites: STATUSPAGE_API_KEY

Install the connector locally

curl -fsSL https://raw.githubusercontent.com/HelpCode-ai/anythingmcp/main/docker-compose.quickstart.yml -o docker-compose.yml
printf 'JWT_SECRET=%s\nENCRYPTION_KEY=%s\n' "$(openssl rand -hex 32)" "$(openssl rand -hex 32)" > .env
docker compose up -d

Open http://localhost:3000/connectors/store, pick Statuspage and paste in the env vars listed under Prerequisites.

Available tools

Example prompts

• "Is anything currently shown as down on our status page? List the open incidents and the components that are not operational."
• "Open an incident "Login errors" with the API component on partial_outage. Show me the text before posting."
• "Post an update on the login incident: the fix is deployed and we are monitoring. Do not notify subscribers."
• "Announce a database maintenance next Tuesday from 22:00 to 23:00 UTC that starts and completes automatically."
• "What was the uptime of each component in the last 90 days, and how many active email subscribers do we have?"

Next steps

• Connect to Claude
• Connect to ChatGPT