Checkly MCP. Monitor API Uptime and Performance Conversationaly

Checkly connects your AI agent right into its monitoring data. You don't gotta open up a dashboard or mess with dashboards—you just talk to it, and your agent handles all the heavy lifting for checking API health and tracking performance. It’s built to let you verify system reliability on demand, whether that means knowing if an endpoint is up or seeing how fast it responds.

Get Full Monitor Inventory: You can start by listing every single API and browser test currently running in your Checkly setup using list_checkly_checks. This gives you a complete inventory of everything configured to check for outages. If you need to know what groups related checks fall into, use list_check_groups to get those major categories.

Audit Account Metadata: Need to know the core details about your workspace? You can pull fundamental account information and organizational structure using get_checkly_account_info. This tool gives you a clear look at how your accounts are set up internally.

List Alerting Destinations: To see where system alerts actually go—like Slack, Email, or PagerDuty—you run list_checkly_alert_channels. That list shows every service that's configured to receive notifications when something breaks.

Manage Background Tasks (Heartbeats): You can check the status of all scheduled background jobs, or 'heartbeat monitors,' by using list_checkly_heartbeats. This confirms that your automated cron processes are actually running and doing their job in the background.

Get Specific Check Details: If you need to know exactly how one specific monitor is set up—say, what parameters it uses or which endpoint it targets—you call get_check_details, specifying the exact API or browser check you're interested in.

Execute Immediate Health Checks: Don’t wait for the schedule. If you gotta verify system health right now, use trigger_check_run. You can manually force any selected monitor to run its test cycle immediately, giving you instant status updates on an endpoint or site.

Retrieve Performance Data: Wanna know how fast things used to be? With get_check_performance_metrics, you pull detailed historical data. This tool gathers performance metrics like average response times for any specific check, letting you track trends over time instead of just seeing if it's up or down.

How It All Works Together: Your agent pulls together all this info naturally. You can start by getting a full list of checks with list_checkly_checks, then select one to run an immediate test using trigger_check_run. After the test finishes, you can feed that check's identifier into get_check_performance_metrics to grab historical data and see how it stacks up against its past performance.

You can cross-reference this by checking if the alert channel list via list_checkly_alert_channels includes PagerDuty, which tells you where system failures will actually be reported.

It’s all about getting specific answers fast. Checking account details with get_checkly_account_info helps set context for what's being monitored, while confirming background jobs are running via list_checkly_heartbeats means the monitoring itself isn't failing silently. You control every step from inventory listing to performance auditing.