2.0.0.0

Features

1. Set built-in AI token TTL to 7 days and update DeepSeek default model to deepseek-v4-flash.
2. The sample data loop write feature now supports configuring historical event windows, allowing flexible definition of event time ranges for diverse testing and simulation scenarios.
3. The Analysis Workbench now features annotation support, allowing users to add notes on charts and data points for enhanced team collaboration and information sharing, improving the readability and traceability of analysis results.
4. Support multi-language prompt display for built-in skills.
5. Upgrade MCP server management to template mode.
6. Sync sidebar tree and breadcrumb on channel detail URL navigation.
7. AI Function page supports direct SKILL file upload.
8. The general element page now supports viewing version history for easy change tracking.
9. Add Agentic AI connection management with /agent API.
10. Actions now support the AI Agent type, enabling AI‑based autonomous decision‑making and automated execution.
11. Analysis workbench supports saving AI insights as panel notes, optimize PDF report export.
12. Add AI Function creation entry from any page with auto-filled context type.
13. Add send_im_message agent tool for delivering Markdown reports to user's bound Feishu/Telegram/WeChat accounts.
14. Add slash skill commands to AI chat with autocomplete and filtering.
15. Panel templates now support custom naming patterns for quick identification and categorized management.
16. Enable file-level editing for user-defined skills.
17. Support global document uploads and tree-based element picker.
18. Add histogram, candlestick, and heatmap panel types.
19. Version control now supports selecting a target version from the version list for rollback, enabling quick restoration to a specific state.
20. Add CLS identity signature and ELS verification with Java SDK support.
21. Split RCA and general analysis, per-module standalone analysis.
22. RCA supports cross-element attribute correlation analysis.
23. Add DingTalk IM channel support with proactive send.
24. The element attribute list page now supports filtering by name, helping you quickly locate target attributes.
25. Click a point on the configuration canvas to quickly view its trend chart for instant data analysis.
26. Move composite metric loop to Go orchestration; remove PDF on annotation delete.
27. The element analysis list now includes a "Include Child Elements" option, enabled by default, for conveniently viewing all analyses associated with child elements.
28. Adapt IDMP MCP to 2026-07-28 protocol and official SDK.
29. The panel's auto-downsampling now supports dynamic adjustment, automatically matching the optimal downsampling granularity during zoom in/out operations for smoother data viewing.
30. When accessing the Analysis Workbench via a shared link, the panel interpretation feature fails to load properly, preventing the interpretation content from being viewed.
31. After deploying All-in-One via the Linux installation package, the real-time analysis module fails to start properly, making the feature unavailable.
32. The anomaly detection feature in the panel's window analysis cannot select a target, preventing configuration from being completed.

Improvements

1. Provide official recommended model parameter configurations.
2. The Analysis Workbench now supports batch addition of events, allowing you to import multiple events at once for improved efficiency and ease of use.
3. Automatically validates the TSDB version required for KPI attributes, tag references, and status windows.
4. Optimize the UI styles of the MCP list, function list, and MCP editing page.
5. Unit names are now required to be unique only within their category, rather than globally, offering greater naming flexibility.
6. Enrich general analysis content, merge general and root cause analysis into one button.
7. The Analysis Workbench now supports copy and paste, enabling fast reuse of configurations and data for improved operational efficiency.
8. Introduce ruff linter for tdmodel and integrate into CI.
9. Multi-lane trend charts now support independent Y-axis settings for each lane, meeting the needs of comparative analysis across different data scales.
10. The bar chart panel now supports custom Y‑axis ranges, allowing you to focus on specific data intervals for more flexible analysis.
11. Show localized provider connection error messages in AI UI.
12. Stabilize AI recommendation and recommendation generation E2E tests.
13. Add knowledge retrieval E2E test with references assertion.
14. AI and model-related pages now adopt a unified visual style, ensuring interface consistency and smooth interaction.
15. Add WeChat enable toggle to channel detail page, fix channel state isolation on type switch.
16. Harden LLM logs with centralized redaction and structured logging.
17. Element information now includes creation time for easy record tracking and change history reference.
18. Downsampling has been removed from the table panel.
19. Optimized the layout of the panel edit page by moving the downsampling button to the top area for more intuitive and convenient operation.
20. AI buttons auto-disabled when connection is unavailable, with tooltip showing reason.
21. Optimized the attribute operation workflow in the Analysis Workbench for improved configuration efficiency and interaction responsiveness.
22. Import and export operations have been upgraded to asynchronous task mode. Task progress and results will be pushed to the frontend in real time via notifications, eliminating the need to wait for page responses and improving the user experience.
23. Optimized child element list expansion performance for faster loading and responsiveness under large data volumes.
24. Refactored the boundary value configuration for panels including trend charts, bar charts, and the Analysis Workbench to enhance flexibility and accuracy in data visualization.
25. The attribute details page now supports viewing and clicking to navigate to the associated attribute template.
26. During resource management operations (delete, update, and add), an unexpected system crash may result in failure to recover automatically, posing a stability risk.
27. Provide multi-platform index recovery scripts to streamline the index recovery process after anomalies and improve troubleshooting efficiency.
28. Thresholds in panel configuration should not be set independently; they should be managed as part of the color scheme to maintain visual consistency.
29. Optimized the paging display of the left tree structure to keep it aligned with sibling elements after page turns, improving browsing consistency.
30. Added a "Share" entry to the more options menu (···) of panels and dashboards for quick link sharing.
31. Sample backend enhancement: configuration data is now cached, significantly improving read performance and system responsiveness.
32. Optimized canvas dragging in configuration editing for smoother interaction and faster response.
33. The "Include Child Elements" option in the element's panel, dashboard, and event lists is now enabled by default, making it easy to view data from all associated child elements together.
34. The attribute list page now uses an action icon to control the display of hidden attributes, making the interface cleaner and the operation more intuitive.
35. The UOM operation interface has been optimized for smoother interaction and improved configuration efficiency.
36. The 6 default sample templates have been fully updated and replaced with new example content.
37. Optimize upload and support more file types (38 types tested).
38. Optimize upload error handling and support more file types.
39. Use LiteParse as default PDF/Office/image-to-Markdown backend for improved document Q&A quality.
40. AI recommended panels not relevant to customer data, optimize element recommendation logic.
41. Optimized Git repository initialization to automatically ignore temporary files from file transactions, preventing accidental commits and keeping the repository clean.
42. AI intent routing and response language stability.
43. Optimized the health detection mechanism from frontend services to backend services, improving service availability monitoring and self‑healing capabilities.

Bug Fixes

1. Add download links for data interpretation and root cause analysis with markdown and PDF support.
2. Add missing frontend i18n configuration in IDMP.
3. In dark mode, the AI Function conversation window occasionally displays in light mode, affecting interface theme consistency.
4. Email server configuration now supports no‑authentication mode, allowing mail delivery without providing a username or password.
5. Force UTF-8 stdio in Python runtime and improve crash observability on Windows.
6. Stub element/template index endpoints as no-ops to fix delete endpoint errors.
7. Hover effects are not triggered for canvas symbols in edit mode, resulting in missing interactive feedback.
8. The event search function is abnormal, displaying chaotic results that fail to accurately match the target events.
9. The window analysis added in the Analysis Workbench cannot be deleted, affecting management flexibility.
10. The change reason field appears incorrectly in the model management page and should not be displayed, affecting interface cleanliness and operational smoothness.
11. State transitions in the configuration may cause display anomalies in symbols, affecting real-time data visualization.
12. Once historical events generated by a real-time analysis are manually deleted, the same analysis instance cannot regenerate them.
13. AI chat now searches user-uploaded documents for non-TDengine topics.
14. After opening a configuration dashboard from the configuration list in the dashboard main menu, clicking edit triggers an error indicating that the corresponding element cannot be found.
15. Multiple attribute fields cannot be edited individually; all modifications must be made via the attribute template.
16. KPI attribute save operation failed.
17. The Y-axis settings in the Analysis Workbench fail to apply, affecting chart data display.
18. When viewing the child element panel of an intermediate element in the element browser, a system error is triggered and the panel fails to load.
19. Clicking any filter item in the element filter and then clicking the top‑left logo triggers a system error, causing abnormal page behavior.
20. The offline documentation page experiences abnormal style rendering when opened, affecting the normal viewing of document content.
21. After canceling the limit conditions of an attribute template, attempting to set them again triggers a system error.
22. After deleting a dashboard from the left tree structure in the dashboard main menu, the right-side list page does not refresh automatically and requires a manual refresh to update the display.
23. When saving a modified attribute template, errors may be triggered in some scenarios, causing the save to fail.
24. Windows startup scripts may be mistakenly deleted by some antivirus software, causing startup failure.
25. Centralize Provider configuration — prevent user-selected LLM provider from being overwritten by locale default on restart; remove user-workspace LLM connections API, use system Agent/Connections instead.
26. An abnormal change in the value type of an element attribute may unintentionally clear the configured data reference attributes.
27. An error occurred during the simple import operation, preventing it from completing successfully.
28. The bar gauge in the panel template fails to render properly.
29. The IDMP service address referenced in the invitation email is incorrect, preventing users from accessing it properly.
30. Explorer license page incorrectly displaying IDMP-related grants and status.
31. OpenAI gpt-5.5 and newer models reporting max_tokens unsupported error.
32. Independent attributes of elements are not currently included in search results and cannot be retrieved during searches.
33. Some entries in the element search results cannot be accessed, showing a “element not found” error.
34. Document upload progress stuck at some percentage.
35. The IDMP service address referenced in the invitation email is incorrect, preventing users from accessing it properly.
36. AI Function panel interpretation not executing as expected.
37. The element does not currently support AI‑recommended real‑time analysis.
38. Panel query response times are excessively long in some environments, affecting data loading efficiency and operational smoothness.
39. AI analysis creation NPE due to missing triggerType.
40. Remove thinking content from Feishu and Telegram responses.
41. Adding events to the Analysis Workbench is slow, affecting operational efficiency.
42. After clicking an attribute in the search results and then clicking the top-left icon, the system triggers an error, interrupting the page operation.
43. Multiple exceptions occurred during the upgrade process, preventing the upgrade from completing successfully.
44. Clicking to view the write task results in a blank screen, and the task details fail to load.
45. An error occurred during data import, causing the import task to fail.
46. Local Qwen model reports AI inaccessible by merging multiple system prompts into one.
47. Custom LLM provider with empty API key reports model connection not configured.
48. System config (collect, crash report, auto-refresh, interval) silently reset when toggling WeChat switch.
49. AI suggestion removal index misalignment causing wrong item deletion.
50. An error occurs when searching for elements in the configuration dashboard, preventing the search from completing successfully.
51. A data backup operation in a container environment triggers an error, causing the backup task to fail.
52. An error occurred during element search, preventing search results from loading properly.
53. After upgrading the Windows All-in-One version, the system prompts that registration is required.
54. License Server TSDB and IDMP expiry times inconsistent.
55. Feishu bot chat message format incorrect.
56. Website chat display format corruption.
57. The status window in real-time analysis does not currently support string‑type status attributes; such configurations will not trigger event generation.
58. Failed to save the modified analysis: the specified table does not exist.
59. IM 64-bit message ID precision loss causes chat history corruption.
60. Explicit Slash Skill execution priority and argument expansion.
61. The support for the DERIVATIVE function in analysis output attributes is abnormal, causing calculation results to deviate from expectations.
62. AI chat session CRUD IDOR - add user ownership validation.
63. SSE subscription escapes component scope causing memory leak.
64. AI streaming response written to wrong session - capture target message per request.
65. SSE Cancellable discarded causing worker thread blocked for 12 minutes.
66. Missing return in AiDocumentService non-admin guard causes permission bypass.
67. Streaming response full re-parse and persistence per frame causes O(n²) lag.
68. The search operation errors out and is interrupted if either the start or end time of the time range is left empty.
69. IM Bot session APIs missing ownership check allows cross-user IDOR.
70. EventMapper SQL injection - add SqlSafetyValidator read-only whitelist.
71. Feedback request aborts in-flight long request, use independent abortController per request.
72. Multiple exceptions occur during version control usage.
73. The query area of the "Child Elements/Attributes" list in the element browser has abnormal styling, affecting interface cleanliness and user experience.
74. Panel interpretation and RCA errors due to search_documents type mismatch and PDF conversion failure.
75. An error occurred during the similarity analysis, preventing the results from being generated.
76. AI chat misreports metric stats as element panels.
77. The example code in the Python SDK documentation does not work as expected in the actual environment, indicating a discrepancy.
78. The REST API call for retrieving a single element via the Python SDK fails, preventing the element data from being retrieved.
79. Editing the configuration dashboard: selecting multiple symbols causes significant lag, preventing drag-and-drop movement.
80. The configuration dashboard does not currently support customizing the background color via RGB values, preventing precise color adjustments.
81. Data export failed because the audit field request_params exceeded the length limit.
82. AI Function button wraps to a new line.
83. During panel sharing, the "Submit" button becomes unresponsive in some scenarios, preventing the sharing operation from completing.
84. The "Comment" field on the element details page has been renamed to "Annotation" to maintain consistent terminology across the platform.
85. The title of the statistics panel in the dashboard is not fully displayed and appears truncated.
86. When sharing a dashboard, the "Submit" button is unresponsive in some scenarios, preventing the sharing operation from completing.
87. After deleting the element tree, the status is updated to "unavailable," but the deletion task does not complete within the expected time, resulting in a failure.
88. The added panel cannot be displayed properly in the configuration dashboard.
89. Case-sensitivity in algorithm matching for natural-language anomaly detection creation.
90. Notification threads may hang silently, blocking all channels.
91. AI analysis creation generates duplicate attributes.
92. Report download returns wrong report after panel interpretation follow-ups.
93. IM binding code brute-force enumeration prevention and collision safety.
94. After a long wait, the data export page eventually shows no data.
95. The creation time field is occasionally missing from the element search result list.
96. AI interpretation of analysis workbench fails on official demo.
97. Select default AI provider by activation locale (zh→DeepSeek, others→OpenAI).
98. The associated virtual table name should remain unchanged when the element name is modified, ensuring independent management and preventing unintended impacts on data mapping.
99. During data asset import, a large number of TSDB query requests accumulate, causing system response delays and reducing import efficiency.
100. Output attributes are empty when creating an analysis for an element without a template.
101. An exception occurs when modifying the KPI attribute, preventing the operation from completing successfully.
102. CSV data model import suffers from a severe performance bottleneck, with an extremely low import rate that takes two days to import just over 700 elements, significantly impacting data initialization efficiency.
103. The sample data loading button is unavailable in some cases, affecting the data initialization process.
104. A system error occurs when opening a panel in a top‑level element, preventing the panel from loading properly.
105. Sample data upload fails, causing the data import process to be interrupted and preventing initialization.
106. After deleting an element, the system does not automatically clear the events previously generated by it, leaving residual event data that deviates from expected behavior.
107. The Excel Add-In displays Chinese characters in its interface even when the language is set to English, affecting the localization experience and interface consistency.
108. After version control is enabled, modifications to real-time analysis cannot be performed successfully, preventing the changes from taking effect.
109. The selectable units for time intervals in the panel's default value filling are incomplete.
110. Failed to add the calculation expression for output attributes when creating a sliding window real-time analysis.
111. Import of sample data dashboard failed; the import operation could not be completed.
112. Data export failed: The system could not locate the export task with the specified ID (xxxx). The task may have been cleared or the ID may be invalid.
113. When converting a dashboard to a template, the sample data download fails, preventing the conversion from completing.
114. Data import failed: the specified panel category was not found. Please check if the category exists or reselect.
115. The statistical values in the Analysis Workbench legend display abnormally, with rendered values inconsistent with actual data.
116. The sharing token for panels and dashboards currently cannot be disabled.
117. In the connection list page, clicking the "Easy Import" button within the "More Actions" menu of each connection yields no response, preventing the import function from being triggered and affecting the data initialization process.
118. The REST API call for retrieving a single element via the Python SDK fails, preventing the element data from being retrieved.
119. The annotation list does not refresh when switching elements, still showing the previous element's content.
120. In full‑screen mode, the time selection button in the configuration popup fails to respond to clicks.
121. In the Add dialog for the attribute list in the Analysis Workbench, an extraneous dot appears after the selection box in the first column, affecting interface cleanliness.
122. In scheduled delivery tasks for panels and dashboards, the backend restarts the browser rendering engine each time it runs, rather than reusing existing instances, which may lead to increased resource overhead.
123. When performing area selection analysis in the Analysis Workbench, the system reports an SQL syntax error, preventing the analysis from completing.
124. After adding content to the Analysis Workbench, the newly opened page loads slowly, affecting operational efficiency and user experience.
125. Page loading may fail due to high browser memory usage.
126. After the upgrade, old panels fail to load properly, which may affect access to existing views.
127. Empty model list when remote provider catalog is unreachable.