1.0.20.0

Features

1. The dashboard editor now supports adding editable rich text components directly, making it easy to embed notes or dynamic content.
2. Analysis now supports adding notification actions to trigger alerts automatically.
3. Add AI document retrieval PageIndex resource package generation.
4. Version control now supports manual release, allowing you to control the timing of version activation for easier review and gradual rollout.
5. The heatmap now supports dimensions other than time (e.g., categorical, numeric) to meet multi-dimensional data distribution analysis needs.
6. The library now supports action template management for unified configuration and reuse.
7. Show panel component directly in AI chat.
8. The Excel plugin supports inserting only data values, automatically ignoring headers and description rows to streamline data writing.
9. Generated events support both single and batch deletion for flexible and efficient event management.
10. The CSV import feature for attribute data now provides a downloadable template for quick and compliant data preparation.
11. AI Telegram Bot integration — receive messages via Telegram Bot API Webhook, reply through ACP Agent. Supports text/image/file messages with optional user whitelist, multi-turn conversation.
12. AI Feishu Bot integration — receive messages via Feishu Open Platform Webhook + Event Subscription, reply through ACP Agent. Supports DM and group @bot mentions, user identity binding via invite code, multi-turn conversation context.
13. Dashboard editing now supports panel grouping for easy categorization and navigation.
14. ACP resource monitoring full implementation — complete system resource monitoring for AI Agent, including real-time CPU/memory/disk usage collection and display, with MCP tool support for resource status queries.
15. ACP domain glossary — provide domain glossary support for AI Agent, enabling it to understand and use professional domain terminology to improve conversation quality and accuracy.
16. Implement basic Agent based on Claude Code with ACP protocol support — Go-based AI Agent backend supporting ACP protocol for external Agent communication, MCP tool calls, session management, and streaming responses.
17. Version control now supports GitHub integration for seamless management of configuration changes with code repositories.
18. Introduce external Agent (Claude Code) as AI service backend — switch AI backend from built-in Python LangGraph to external Agent via ACP protocol, supporting MCP tool calls, session management, multi-platform IM integration (Feishu/Telegram/WeChat).
19. The analysis panel now supports saving the current analysis state, making it easy to resume and continue your analysis at any time.
20. Full backup is supported, allowing one-click complete backup of all data to ensure business recoverability.
21. Supports MFA/2FA (multi-factor authentication) to enhance account login security.
22. User password policy has been enhanced to support custom complexity and expiration, strengthening account security.
23. Add a configurable main menu for Quality Analysis, supporting flexible display/hide options as needed.
24. Element and event templates now support custom small icons for quick identification.
25. Release IDMP cli skills and claude code plugin.
26. Convert existing AI features to Skills — standardize IDMP's existing AI capabilities (panel creation, recommendation, RCA, analysis recommendation) as Skills for on-demand loading by ACP Agent.
27. Cloud service supports TSDB/IDMP MCP — cloud service integrates TSDB and IDMP MCP tool services, enabling external AI Agents to access TDengine database and IDMP platform capabilities via MCP protocol.
28. Integrated offline IDMP documentation for easy reference anytime, even without an internet connection.
29. New version of ai chat.
30. AI chat integrate with Feishu agent bot.
31. Add Agentic AI Library pages (Skills/MCPs list) and User Profile Agent tab.
32. IM binding list always shows all 3 channels with status; Telegram atomic binding flow + unbind cache cleanup.

Improvements

1. Added a toggle in basic system configuration to enable/disable the Quality Analysis menu in the admin backend.
2. The data writing menu has been moved under the corresponding TDengine connection node, aligning data write operations more closely with connection management.
3. Improved refresh behavior for attribute template details page: the page remains on the details view after refresh instead of returning to the list.
4. Remove non-business default Path attributes from samples to better reflect real-world business scenarios.
5. Use global locks to ensure atomicity and data consistency for merge and rollback operations in version control.
6. Added global resource lock mechanism to support multi-user collaboration and prevent concurrent conflicts.
7. ACP init fault tolerance — split app/__init__.py try/except into two levels: config/ImportError cause startup failure (hard error), MCP init exceptions only trigger downgrade warning (soft degradation), preventing single MCP failure from crashing the entire AI service.
8. Data model import tasks now support asynchronous workflow notifications, allowing real-time tracking of import progress and results.
9. Rename API key UI text from key to API Key.
10. Optimized attribute value query performance for faster retrieval and reduced latency.
11. Optimize panel interpretation and RCA display format.
12. The analysis module has been upgraded to Real-Time Analysis, and the Analysis Panel has been renamed to Analysis Workbench, offering a more real-time and integrated data analysis experience.
13. Attribute names now support the plus sign (+) for more flexible naming.
14. The Analysis Workbench now supports sharing, allowing you to quickly share analysis results with team members and improve collaboration efficiency.
15. Support token usage statistics.
16. Improved loading performance for the element panel card list, enabling faster page rendering and a smoother browsing experience.
17. Optimized query performance for statistical values of element metric attributes, significantly reducing response latency and improving data retrieval efficiency.
18. Speedup ai chat response time costs and enhance quality.
19. Opt: add time range to AI-generated SQL to avoid full table scan.
20. AI-recommended SQL causes tsdb backlog, optimize SQL generation logic.

Bug Fixes

1. Add max_completion_tokens parameter to LLM API calls, fixing interaction failures with privately deployed LLMs.
2. Route AI visualization requests to create_panel instead of execute_code/matplotlib, support start_time/end_time time window for panels.
3. The diff function is not supported in counting window analysis.
4. After uninstalling the sample, re-uploading and loading triggers a system error indicating that the element template already exists, causing the loading operation to fail.
5. The upgrade from 1.0.19.2 to 1.0.20.0 was not completed successfully.
6. After enabling version control, the associated submenu does not appear properly, which may affect branch management and version switching operations.
7. After configuring quality bit filter conditions in the panel, the filtering results do not take effect correctly, which is inconsistent with expectations.
8. An error appears in the browser console when opening the panel, which may affect normal loading and interaction.
9. The timeline in the analysis panel lacks color differentiation, making it difficult to quickly distinguish between different events or states.
10. When importing data models via CSV, currently only the first row of each element template can be imported; the remaining rows cannot be written successfully.
11. AI service compatibility issues.
12. Opening the offline help documentation now redirects to the system home page, preventing direct access to the documentation content.
13. Clicking after an attribute search triggers a system error, affecting the normal workflow.
14. Import failed: the specified attribute could not be found. Please verify the import file or template configuration.
15. When creating a panel, selecting a metric that includes a quality bit for querying triggers a system error, causing the panel creation to fail.
16. The scheduled backup task triggers an exception during system startup, causing startup failure and affecting system availability.
17. Multiple yellow warning windows frequently pop up in the Analysis Workbench, disrupting the normal workflow.
18. When a panel is configured with multiple metrics, duplicate query fields may appear, potentially causing data anomalies or degraded query performance.
19. After attribute binding configuration, the symbol value in the configuration interface fails to refresh in real time, causing data display delays that affect monitoring effectiveness.
20. Parameter display error when calling REST API via OpenAPI in sample data.
21. New user invitation email sent successfully, but registration activation failed.
22. Panel interpretation refresh incorrectly reused session memory instead of starting fresh.
23. Idmp-easyuse skill panel creation uses description as title unexpectedly.
24. The prediction configuration in the Analysis Workbench fails to automatically load the previously saved settings when reopened.
25. Alert filter natural language generation issue.
26. Composite metric generation error.