Dynatrace MCP Server

# @dynatrace-oss/dynatrace-mcp-server ## Unreleased Changes ## 0.8.0 ### Tools - Added a `limit` argument to the `get_kubernetes_events` tool, allowing you to control the number of events returned and improving performance for large clusters ### Other Changes - Fixed some typos - Respond with a proper JSON RPC Error message ## 0.7.0 **Highlights** 🔒 Human approval for critical operations 🔍 Enhanced entity discovery with automatic detection 🛠️ Improved error handling and internal optimizations ### Tools - Removed the `get_entity_details` tool and consolidated its functionality into the `find_entity_by_name` tool for a streamlined user experience - Enhanced the `find_entity_by_name` tool with automatic entity name detection for improved usability - Added human approval steps for critical operations in `send_email`, `send_slack_message`, `create_workflow_for_notification`, and `make_workflow_public` tools to ensure user consent and prevent unintended actions ### Other Changes - Disabled Grail budget enforcement for Dynatrace-internal development and hardening stages to facilitate testing and development workflows - Improved error handling for environments without Davis Copilot enabled, now providing direct links to enable the feature ## 0.6.1 - Fixed an issue with MCP communication failing with `SyntaxError: Unexpected token 'd'` due to `dotenv` - Added Support for Google Gemini CLI ## 0.6.0 **Highlights**: 💰 Grail budget tracking and cost control 📧 Send findings via E-Mail via the Dynatrace E-Mail API 🔧 Enhanced tool annotations for better LLM integration 🏪 Published to official MCP Registry and GitHub MCP Registry ### Scopes - Added OAuth scope `email:emails:send` to enable email functionality ### Tools Added/Removed - Added `send_email` tool for sending emails via the Dynatrace Email API with support for multiple recipients (TO, CC, BCC), custom subject lines, and rich body content - Added tool-annotations `readOnlyHint`, `idempotentHint`, and `openWorldHint` to improve tool usage by providing better hints to LLM clients about tool behavior - Added next-steps guidance to `get_entity_details` tool to help users discover related metrics, problems, and logs for entities ### Other Changes - Fixed an issue with the stateless HTTP server that prevented it from accepting multiple simultaneous connections - Added Grail budget tracking with `DT_GRAIL_QUERY_BUDGET_GB` environment variable (default: 1000 GB, setting it to `-1` disables budget tracking), providing cost control and visibility with warnings and alerts in `execute_dql` tool responses - Added budget enforcement that prevents further DQL query execution when the configured Grail budget has been exceeded, protecting against unexpected costs - Improved Davis CoPilot integration by migrating to the official `@dynatrace-sdk/client-davis-copilot` package, enhancing reliability and maintainability while reducing manual API implementation - Added metadata output to `execute_dql` tool which includes scanned bytes information, enabling better cost tracking for Dynatrace Grail data access - Added telemetry via Dynatrace OpenKit to improve the product with anonymous usage statistics and error information, enhancing product development while respecting user privacy (can be disabled via `DT_MCP_DISABLE_TELEMETRY` environment variable) - Added `server.json` configuration and published the MCP server to the official MCP Registry, making it easier for users to discover and install the server ## 0.6.0 (Release Candidate 2) - Fixed an issue with the stateless HTTP server that prevented it from accepting multiple simultaneous connections - Added Grail budget tracking with `DT_GRAIL_QUERY_BUDGET_GB` environment variable (default: 1000 GB, setting it to `-1` disables budget tracking), providing cost control and visibility with warnings and alerts in `execute_dql` tool responses - Added budget enforcement that prevents further DQL query execution when the configured Grail budget has been exceeded, protecting against unexpected costs - Added `send_email` tool for sending emails via the Dynatrace Email API with support for multiple recipients (TO, CC, BCC), custom subject lines, and rich body content - Added OAuth scope `email:emails:send` to enable email functionality - Improved Davis CoPilot integration by migrating to the official `@dynatrace-sdk/client-davis-copilot` package, enhancing reliability and maintainability while reducing manual API implementation ## 0.6.0 (Release Candidate 1) - Added metadata output to `execute_dql` tool which includes scanned bytes information, enabling better cost tracking for Dynatrace Grail data access - Added next-steps guidance to `get_entity_details` tool to help users discover related metrics, problems, and logs for entities - Added telemetry via Dynatrace OpenKit to improve the product with anonymous usage statistics and error information, enhancing product development while respecting user privacy (can be disabled via `DT_MCP_DISABLE_TELEMETRY` environment variable) - Added `server.json` configuration and published the MCP server to the official MCP Registry, making it easier for users to discover and install the server - Added metadata output which includes Grail scanned bytes (for cost tracking) to `execute_dql` - Added next-steps for `get_entity_details` to find out about metrics, problems and logs - Added Telemetry via Dynatrace OpenKit to improve the product with anonymous usage statistics and error information (can be disabled via `DT_MCP_DISABLE_TELEMETRY` environment variable) ## 0.5.0 **Highlights**: 🚀 Davis CoPilot AI, supporting natural language to DQL 🌐 HTTP transport support 🔑 Platform Token authentication 📚 Tool consolidation into `execute_dql` ### Scopes - Removed unnecessary scope `environment-api:security.problems:read` as it's no longer needed - Removed unneeded scopes `environment-api:slo:read` and `environment-api:metrics:read` as functionality is handled via the `execute_dql` tool ### Tools Added/Removed - Added tools to translate between natural language and DQL via Davis CoPilot, enabling easier query creation - Added tool to chat with Davis CoPilot for interactive assistance and guidance - Removed `get_logs_for_entity` tool in favor of the more flexible `execute_dql` tool - Removed `get_vulnerability_details` tool as the same functionality can now be achieved with a simple `execute_dql` call, simplifying the tool set - Removed `get_problem_details` tool as the same functionality can be achieved with a simple `execute_dql` call ### Other Changes - Added cost considerations disclaimer in README about Dynatrace Grail data access to help users understand potential costs - Added `dtClientContext` to `execute_dql` tool, enabling usage monitoring for Grail access and better cost tracking - Added information about Semantic Dictionary for `execute_dql` tool description, improving user guidance for DQL queries - Added Streamable HTTP transport support with `--http`/`--server`, `--port`, and `--host` arguments, enabling you to run the server over HTTP while maintaining stdio as the default for backward compatibility - Enhanced `find_entity_by_name` tool to include all entities from the Smartscape topology, providing comprehensive entity discovery capabilities - Optimized `get_monitored_entity_details` tool to use direct entity type lookup for better performance and faster response times - Improved `list_vulnerabilities` tool to use DQL statements instead of classic API, aligned parameters with `list_problems` tool for consistent user experience - Added comprehensive AI-Powered Observability Workshop Rules with hierarchical workflow architecture for advanced analysis scenarios - Enhanced README with advanced analysis capabilities including incident response, security compliance, and DevOps automation workflows - Added support for multi-phase incident investigation, cross-data source correlation, and precise root cause identification - Introduced streamlined rule structure optimized for LLM context windows with all files under 6,500 tokens for better AI assistant performance - Added integration guides for multiple AI assistants including Amazon Q, Cursor, Windsurf, Cline, and GitHub Copilot - Enhanced example prompts with sophisticated use cases for transaction analysis, security assessment, and DevOps workflows - Removed `metrics` from `execute_dql` example with `fetch` to improve clarity - Clarified usage of `verify_dql` to avoid unnecessary tool calls and improve efficiency - Improved `list_problems` tool to use DQL statements for retrieving data from Dynatrace and provide better next steps for problem resolution - Added support for authorization via Platform Tokens using the `DT_PLATFORM_TOKEN` environment variable, providing an alternative authentication method ## 0.5.0 (Release Candidate 4) - Added Streamable HTTP transport support with `--http`/`--server`, `--port`, and `--host` arguments (default remains stdio for backward compatibility) - Adapted `find_entity_by_name` tool to include all entities from the Smartscape topology. - Optimized `get_monitored_entity_details` tool to use direct entity type lookup for better performance. ## 0.5.0 (Release Candidate 3) - Improved `list_vulnerabilities` tool to use DQL statement instead of classic API, and aligned parameters with `list_problems` tool - Removed `get_vulnerability_details` tool as the same can now be achieved with a simple `execute_dql` call - Removed scope `environment-api:security.problems:read` as it's no longer needed - Added comprehensive AI-Powered Observability Workshop Rules with hierarchical workflow architecture - Enhanced README with advanced analysis capabilities including incident response, security compliance, and DevOps automation - Added support for multi-phase incident investigation, cross-data source correlation, and precise root cause identification - Introduced streamlined rule structure optimized for LLM context windows (all files under 6,500 tokens) - Added integration guides for multiple AI assistants (Amazon Q, Cursor, Windsurf, Cline, GitHub Copilot) - Enhanced example prompts with sophisticated use cases for transaction analysis, security assessment, and DevOps workflows - Removed unneeded scopes `environment-api:slo:read` (no tool is using this) and `environment-api:metrics:read` (anyway handled via execute DQL tool) - Removed `metrics` from `execute_dql` example with `fetch`. - Clarified usage of `verify_dql` to avoid unnecessary tool calls. ## 0.5.0 (Release Candidate 2) - Improved `list_problems` tool to use a DQL statement to retrieve data from Dynatrace, and provide better next steps - Removed `get_problem_details` tool, as the same can be achieved with a simple "execute_dql" call - Removed scope `environment-api:problems:read` as it's no longer needed ## 0.5.0 (Release Candidate 1) - Added support for Authorization via Platform Tokens via environment variable `DT_PLATFORM_TOKEN` - Added tools to translate between natural language and DQL via Davis CoPilot - Added tool to chat with Davis CoPilot ## 0.4.0 - Improve Authentication - fine-grained OAuth calls per tool - Fixed: Missing scope `storage:security.events:read` for execute DQL ## 0.3.0 - Provide version of dynatrace-mcp-server on startup - Define HTTP user-agent of dynatrace-mcp-server ## 0.2.0 - Added new tool `get_entity_by_name` which allows to find the entity ID of a monitored entity by its name - Improved handling and description of `execute_dql` tool - Improved checking for Dynatrace Environment URL ## 0.1.4 - Improved error-handling of authentication mechanism ## 0.1.3 - Improved error-handling of authentication mechanism ## 0.1.2 - Fix: Added missing `storage:events:read` scope ## 0.1.1 - Maintenance release ## 0.1.0 - Initial Release