# Cronitor MCP for AI Agents MCP > Cronitor brings full visibility into your infrastructure's background jobs, heartbeats, and website uptime. Connect this MCP to instantly track performance metrics, manage alerts, or diagnose failures across any service without switching tabs. Your AI agent acts like a constant SRE assistant right in your chat window. ## Overview - **Category:** developer-tools - **Price:** Free - **Tags:** cron-jobs, uptime-monitoring, heartbeats, incident-management, telemetry ## Description You shouldn't have to jump between ten different dashboards just to see if your nightly backups ran, or why your primary API endpoint is suddenly flaky. This MCP lets you connect Cronitor directly to your preferred AI client, giving you full visibility into your entire infrastructure's health through natural conversation. You can ask your agent to check the status of a critical cron job, query performance metrics for specific sites, or even schedule maintenance windows when planned downtime hits. This capability is huge. Instead of manually running checks, your AI agent handles the monitoring lifecycle: it monitors background jobs, tracks heartbeats, and analyzes site analytics—all from one place. Because this MCP is hosted on Vinkius, you connect once to access Cronitor alongside thousands of other developer tools, making it a single source for operational intelligence. ## Tools ### bulk_issues Perform bulk actions on multiple reported issues at once. ### clone_monitor Create a new monitor using the settings from an existing one. ### create_api_key Generate and issue a new API key for system access. ### create_status_page_component Build individual sections or components that go onto your main status page. ### create_environment Define a new monitoring environment (e.g., staging, production). ### create_group Assemble related monitors into a single logical group. ### create_issue Log and report a new active incident or issue for tracking. ### create_maintenance_window Schedule an official window of planned downtime to prevent alerts from firing unnecessarily. ### create_monitor Set up a new monitor for a specific cron job, heartbeat, or website endpoint. ### create_notification Establish a list of recipients who need to be alerted when an issue occurs. ### create_site Set up a new Real User Monitoring (RUM) site for performance tracking. ### create_status_page Initialize and create an entirely new public status page. ### delete_api_key Revoke and delete a specific API key when it's no longer needed. ### delete_status_page_component Remove status page components that are outdated or incorrect. ### delete_group Delete an entire monitor group from the system. ### delete_monitor Take down and delete a specific, individual monitor. ### delete_site Remove a Real User Monitoring site when monitoring is complete. ### delete_status_page Archive and delete an entire status page instance. ### get_aggregates Retrieve rolled-up aggregates ### get_metrics Requires at least one field parameter. Retrieve performance metrics ### get_monitor Retrieve a specific monitor ### get_site_errors Get site errors ### list_api_keys View all API keys currently associated with the account. ### list_status_page_components Retrieve a list of existing components on status pages. ### list_environments View all defined monitoring environments (e.g., dev, prod). ### list_groups See a list of organized monitor groups. ### list_issues Retrieve a history of all open and closed incidents or issues. ### list_monitors Get an overview, or search for specific existing monitors. ### list_notifications List notification lists ### list_sites List RUM sites ### list_status_pages List status pages ### pause_group Bulk pause or resume a group of monitors ### pause_monitor Use 0 to resume. Pause or resume a monitor ### query_site_analytics Query site analytics ### send_telemetry link. Send telemetry events (pings, metrics) ### update_api_key Update an API key ### update_status_page_component Update status page components ### update_environment Update an environment ### update_group Update a monitor group ### update_issue Update an issue state ### update_monitors Create or update monitors in bulk ### update_notification Update a notification list ### update_site Update RUM sites ### update_status_page Update status pages ## Prompt Examples **Prompt:** ``` What's the status of my nightly database backup job? And also, what were the site error rates yesterday? ``` **Response:** ``` **System Health Check:** * **Database Backup (cron):** Status: **FAILED**. Last run: 2026-06-12. Error Code: `DB_TIMEOUT`. Action needed. * **Website Errors (RUM):** Yesterday's aggregate site errors were 45/10,000 page views. This is a 3% increase over the weekly average. ``` **Prompt:** ``` Can you pause all monitoring for our primary API services while we deploy new code? ``` **Response:** ``` **Action Confirmed:** The 'Primary Services Group' has been successfully paused for the next four hours (until 14:00 UTC). All associated monitors are now suspended, and no alerts will be generated. You can resume monitoring by asking me to `resume group: Primary Services Group`. ``` **Prompt:** ``` List all active monitors and tell me which ones are healthy but haven't reported in 12 hours. ``` **Response:** ``` I checked the list of 34 monitors. Here’s what I found: * ✅ **Healthy/Stale:** `legacy_heartbeat` (Last ping: 15+ hrs). *Recommendation: Check if this service still exists.* * ⚠️ **Needs Review:** `third_party_api` (Status: OK, but last metric pull was 12 hours ago). Monitor is fine, but data is stale. ``` ## Capabilities ### Monitor Infrastructure Status Set up, list, and update monitors for cron jobs, heartbeats, and website health checks across multiple services. ### Gather Performance Metrics Query aggregated data and site analytics to understand latency, success rates, and overall system performance trends. ### Manage Incidents and Alerts Track active issues, create maintenance windows, or manage status pages directly through your AI agent. ### Send Live Telemetry Pings Send job state updates (like 'running', 'complete', or 'failed') to track execution lifecycles in real time. ### Organize Monitoring Assets Group monitors, manage notification lists, and control which environments are being monitored. ## Use Cases ### Investigating a sudden API failure A backend developer notices an endpoint failing. They prompt their agent to check the status by calling `get_monitor` and then use `send_telemetry` to ping the service immediately, getting confirmation of the execution cycle right in the chat. ### Preparing for planned deployment downtime The SRE lead needs to deploy a major system update. They ask their agent to first use `create_maintenance_window` and then pause all related services using `pause_group`, ensuring no alerts fire during the rollout. ### Auditing service performance trends An operations lead needs proof of system stability. They ask their agent to use `get_aggregates` and `list_sites` to pull up historical data on site latency, providing a single dashboard view for stakeholders. ### Scaling monitoring setup A DevOps team needs to monitor five similar services. They prompt their agent to use the `clone_monitor` tool and then run `update_monitors` in bulk, setting up consistency across all new deployments. ## Benefits - Stop manually checking job statuses. Use the `list_monitors` tool to instantly check if a critical background process is healthy or failing. - Get immediate incident visibility using `list_issues`. Instead of searching through logs, your agent pulls up a live list of all active failures and incidents. - Control planned downtime without panic. Schedule maintenance windows with `create_maintenance_window` so alerts automatically silence during deployments. - Improve performance diagnostics by calling `get_metrics` to pull aggregated data on latency across different services in one prompt. - Maintain clean infrastructure by using `delete_monitor` or `delete_group` when a service is deprecated, keeping your monitoring setup accurate. ## How It Works The bottom line is you talk to your agent like a teammate and get immediate operational status reports for your entire stack. 1. Subscribe to the Cronitor MCP and provide your unique API key. 2. Connect this MCP to your AI client (like Claude or Cursor). 3. Use natural language prompts to ask about monitor statuses, query performance data, or trigger incident actions. ## Frequently Asked Questions **How do I use the Cronitor MCP to check if my cron jobs are actually running?** You simply ask your agent for a status report. It checks all defined monitors and tells you which job is healthy, failing, or paused. This saves you from manually checking multiple dashboards just to confirm execution. **Can the Cronitor MCP help me manage my website downtime information?** Yes. You can create a dedicated status page and use the MCP to post updates about ongoing incidents or scheduled maintenance windows, keeping your users informed automatically. **What if I need to track performance metrics for different times of day using Cronitor?** The MCP lets you query aggregated data over specific time ranges. You can pull P50 duration or success rates for the last 24 hours, giving you historical context on system load. **Is Cronitor good for tracking API endpoints that fail intermittently?** Absolutely. By setting up a dedicated monitor and using telemetry pings, your agent tracks the entire execution lifecycle—from running to failure—giving you precise data on intermittent issues. **Does Cronitor help me organize my monitoring setup if I have many services?** Yes. You can use monitor groups to logically cluster related jobs (like 'Payment Services' or 'User Auth'), letting your agent manage them all with one command. **What is the difference between using Cronitor for job monitoring versus website uptime?** Job monitoring tracks background processes that run on a schedule. Website uptime monitors track external-facing endpoints, ensuring the actual pages your users visit are loading correctly and quickly.