Dynatrace MCP. Analyze system health and track metrics from chat.

This server hooks your AI client directly into Dynatrace’s full Environment API (v2). You're not just reading data; you're running deep observability workflows right from the chat or IDE. This gives your agent the power to manage everything—from user accounts and access policies all the way up to triggering performance tests, so you never have to leave your workspace.

Monitoring Status and Problem Management

When something breaks, you need fast answers. You can list all current problems using list_problems, getting a rundown of every open or closed issue in the environment. If you pinpoint the root cause, you've got the ability to mark that system problem as resolved by calling close_problem. For deeper dives, fetching specific details about one incident is simple with get_problem.

You can also track performance trends and request data points for any service or host metric using query_metrics.

Data Ingestion and Discovery

Sometimes the data you need isn't already in Dynatrace. You can push raw operational metrics into the system using the standard line protocol via ingest_metrics, and if you have custom events, you send those straight over with ingest_events. To figure out what's being monitored, you list all entities—hosts, services, apps—by running list_entities.

You can also get a full rundown of available metrics to know exactly what you can query by checking the output of list_metrics.

Synthetic Testing and Topology Mapping

To map out service performance from outside the network, you've got synthetic monitoring. You start by adding a new geographical point using create_synthetic_location, or modifying an existing one with update_synthetic_location. Once you have a location, you deploy a monitor to test endpoints with create_synthetic_monitor, and if things change, updating the schedule or target URL is handled by update_synthetic_monitor.

You can force all configured monitors to run right now using trigger_synthetic_batch for immediate results. When you're done with testing, you can delete either a location (delete_synthetic_location) or an entire monitor configuration (delete_synthetic_monitor). To review past performance runs, you check the detailed results via list_synthetic_executions, and to see what monitors are active, use list_synthetic_monitors.

Infrastructure Management and Configuration

Managing the environment means handling users, permissions, and dashboards. You can add a brand-new user profile with create_account_user, or list every existing user and their status using list_account_users. If you need to change who can access what, you can view all current policies via list_account_policies and then modify permissions for an entire group of users with update_account_group_permissions.

To build out monitoring scope, you create or alter custom devices (entities) using create_custom_entity, and you can also expand your visibility by getting the setup details for anomaly detection across apps (get_anomaly_detection_apps), hosts (get_anomaly_detection_hosts), or specific services (get_anomaly_detection_services).

You control what people see and how it looks. You can build an entirely new dashboard view with create_dashboard, or modify the layout of a pre-existing one using update_dashboard. If you mess up, there's no sweat; you can wipe out a dashboard completely with delete_dashboard. For policy management, you can also list all predefined user groups via list_account_groups and build entirely new custom access rules using create_account_policy.

On the setup side, you view configuration objects with list_settings_objects, check available data structures with list_settings_schemas, or see what calculated service metrics are in play by calling get_calculated_metrics_service.

Cleanup and Visibility

Finally, when you're done managing the environment, you can clean up. You list all currently existing dashboards using list_dashboards. If your monitoring scope changes, listing entity types (list_entity_types) shows you every kind of thing Dynatrace monitors. For a full picture of what's going on, you can view historical data by checking the list of custom events that occurred over time with list_events.