CastrelCastrel AI
Monday, September 14, 2026 · v1.27.0

Dashboard Generation and Execution Upgrade

Dashboard Generation and Execution Upgrade
This release adds LLM-assisted dashboard query generation, and generated dashboards can now be executed as a whole or by group within a task to cut down repeated tool calls and speed up analysis. It also strengthens JumpServer integration permissions and result retrieval, and fixes issues with Feishu Docs installation, knowledge naming conflicts, and custom MCP calls.

New Feature: Dashboard Generation and Execution Upgrade

You can now describe what you need in plain language and let the model generate dashboard query statements for you, building a monitoring view quickly without writing complex query syntax by hand.

LLM-generated dashboard query
LLM-generated dashboard query

Generated dashboards can be executed directly in later tasks: a single call can run an entire dashboard or one of its groups, pulling data from multiple charts at once instead of calling a tool for each chart separately. This cuts down repeated tool calls and helps analysis reach a conclusion faster.

Executing a dashboard within a task
Executing a dashboard within a task

Dashboards often need further refinement after they're generated. Grouping, chart variables, and panel layout can be edited directly in the UI — grouping dashboards by business line, environment, or team, reusing the same template across different services or instances, and resizing or repositioning panels as needed.

1 Enhanced Integration

  • Enhanced: JumpServer (separates admin and user tokens, and lets regular users create jobs and correctly retrieve execution results)

Full Release Checklist

feat

  • Added LLM-assisted dashboard query generation, with the ability to execute an entire dashboard or one of its groups in a task to reduce tool calls.
  • Added dashboard grouping, chart variable substitution, and drag-to-resize layouts as ways to refine and maintain dashboards.
  • Improved JumpServer integration by separating admin and user tokens, and enabling regular users to create jobs and correctly retrieve execution results.

bugfix

  • Fixed a 422 authorization error that blocked Feishu Docs installation due to a missing configuration.
  • Fixed an issue where marking an artifact as knowledge skipped name validation, causing existing knowledge with the same name to be overwritten.
  • Fixed an issue where custom MCP calls incorrectly passed through the connection ID, causing tool calls to fail.