dify1.11.1 Although workflow-runs generate output answers, the conversation lacks such responses. #21200

Closed
opened 2026-02-21 20:11:24 -05:00 by yindo · 4 comments
Owner

Originally created by @zscAZA on GitHub (Dec 21, 2025).

Self Checks

  • I have read the Contributing Guide and Language Policy.
  • This is only for bug report, if you would like to ask a question, please head to Discussions.
  • I have searched for existing issues search for existing issues, including closed ones.
  • I confirm that I am using English to submit this report, otherwise it will be closed.
  • 【中文用户 & Non English User】请使用英语提交,否则会被关闭 :)
  • Please do not modify this template :) and fill in all the required fields.

Dify version

1.11.1

Cloud or Self Hosted

Self Hosted (Docker)

Steps to reproduce

Image

✔️ Expected Behavior

both workflow-runs and conversation components can normally generate and display output content

Actual Behavior

Although workflow-runs generate output answers, the conversation lacks such responses.

Originally created by @zscAZA on GitHub (Dec 21, 2025). ### Self Checks - [x] I have read the [Contributing Guide](https://github.com/langgenius/dify/blob/main/CONTRIBUTING.md) and [Language Policy](https://github.com/langgenius/dify/issues/1542). - [x] This is only for bug report, if you would like to ask a question, please head to [Discussions](https://github.com/langgenius/dify/discussions/categories/general). - [x] I have searched for existing issues [search for existing issues](https://github.com/langgenius/dify/issues), including closed ones. - [x] I confirm that I am using English to submit this report, otherwise it will be closed. - [x] 【中文用户 & Non English User】请使用英语提交,否则会被关闭 :) - [x] Please do not modify this template :) and fill in all the required fields. ### Dify version 1.11.1 ### Cloud or Self Hosted Self Hosted (Docker) ### Steps to reproduce <img width="1904" height="816" alt="Image" src="https://github.com/user-attachments/assets/4c138baf-e65a-4134-b2e6-ed4b88468908" /> ### ✔️ Expected Behavior both workflow-runs and conversation components can normally generate and display output content ### ❌ Actual Behavior Although workflow-runs generate output answers, the conversation lacks such responses.
yindo added the 🐞 bug label 2026-02-21 20:11:24 -05:00
yindo closed this issue 2026-02-21 20:11:24 -05:00
Author
Owner

@dosubot[bot] commented on GitHub (Dec 21, 2025):

The most likely causes for workflow-run outputs not appearing in the conversation component are:

  1. The backend only saves and exposes answers to the conversation UI if the workflow run completes normally and fires the QueueAdvancedChatMessageEndEvent. If the workflow is stopped early (by moderation, annotation reply, manual stop, or error), this event may not fire, and the answer won't be saved or shown in the conversation—even if the workflow-run itself produced output. Recent fixes ensure that all stop endpoints properly record workflow status as STOPPED, but if your deployment is missing these, workflows may get stuck in a running state and not finalize the answer field as expected (details).

  2. There is a known race condition in the output moderation pipeline: if moderation threads finish after the main event loop ends, moderation events (and possible answer updates) can be missed, leading to empty or missing responses in the UI. The suggested fix is to delay queue termination until all moderation threads complete (details).

  3. In self-hosted Docker environments, the default Nginx configuration often blocks Server-Sent Events (SSE), causing streaming output and conversation display issues. To fix this, update your Nginx config for relevant endpoints (like /api and /console/api) to use HTTP/1.1, disable proxy buffering, and set high proxy_read_timeout and proxy_send_timeout values (e.g., 3600s). Avoid overlapping location blocks for the same endpoint, and make sure HTTP/2 is not enabled for SSE endpoints (details) (nginx config advice).

To troubleshoot:

  • Test with different input types (plain text, various file uploads) to see if the issue is input-specific.
  • Check backend logs for abnormal workflow stops, moderation events, or errors that could prevent the answer from being saved.
  • Review your Nginx configuration as described above.
  • Make sure your deployment includes all recent fixes for workflow stop handling and moderation race conditions.

If the problem persists, capture a full workflow run (backend logs, API responses, proxy logs) for further analysis.

To reply, just mention @dosu.


How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other

Chat with Dosu Open in Cursor Join Discord Share on X

@dosubot[bot] commented on GitHub (Dec 21, 2025): <!-- Answer --> The most likely causes for workflow-run outputs not appearing in the conversation component are: 1. The backend only saves and exposes answers to the conversation UI if the workflow run completes normally and fires the `QueueAdvancedChatMessageEndEvent`. If the workflow is stopped early (by moderation, annotation reply, manual stop, or error), this event may not fire, and the answer won't be saved or shown in the conversation—even if the workflow-run itself produced output. Recent fixes ensure that all stop endpoints properly record workflow status as STOPPED, but if your deployment is missing these, workflows may get stuck in a running state and not finalize the answer field as expected [(details)](https://github.com/langgenius/dify/pull/27803). 2. There is a known race condition in the output moderation pipeline: if moderation threads finish after the main event loop ends, moderation events (and possible answer updates) can be missed, leading to empty or missing responses in the UI. The suggested fix is to delay queue termination until all moderation threads complete [(details)](https://github.com/langgenius/dify/issues/29033). 3. In self-hosted Docker environments, the default Nginx configuration often blocks Server-Sent Events (SSE), causing streaming output and conversation display issues. To fix this, update your Nginx config for relevant endpoints (like `/api` and `/console/api`) to use HTTP/1.1, disable proxy buffering, and set high proxy_read_timeout and proxy_send_timeout values (e.g., 3600s). Avoid overlapping location blocks for the same endpoint, and make sure HTTP/2 is not enabled for SSE endpoints [(details)](https://github.com/langgenius/dify/issues/19324) [(nginx config advice)](https://github.com/langgenius/dify/issues/24848). To troubleshoot: - Test with different input types (plain text, various file uploads) to see if the issue is input-specific. - Check backend logs for abnormal workflow stops, moderation events, or errors that could prevent the answer from being saved. - Review your Nginx configuration as described above. - Make sure your deployment includes all recent fixes for workflow stop handling and moderation race conditions. If the problem persists, capture a full workflow run (backend logs, API responses, proxy logs) for further analysis. <!-- Dosu Comment Footer --> *To reply, just mention [@dosu](https://go.dosu.dev/dosubot).* --- <sup>How did I do? [Good](https://app.dosu.dev/response-feedback/ca29ad17-93a3-4a12-85c9-70e99ca8c9b3?feedback_type=great_response) | [Irrelevant](https://app.dosu.dev/response-feedback/ca29ad17-93a3-4a12-85c9-70e99ca8c9b3?feedback_type=irrelevant_answer) | [Incorrect](https://app.dosu.dev/response-feedback/ca29ad17-93a3-4a12-85c9-70e99ca8c9b3?feedback_type=incorrect_sources) | [Verbose](https://app.dosu.dev/response-feedback/ca29ad17-93a3-4a12-85c9-70e99ca8c9b3?feedback_type=too_verbose) | [Hallucination](https://app.dosu.dev/response-feedback/ca29ad17-93a3-4a12-85c9-70e99ca8c9b3?feedback_type=hallucination) | [Report 🐛](https://app.dosu.dev/response-feedback/ca29ad17-93a3-4a12-85c9-70e99ca8c9b3?feedback_type=bug_report) | [Other](https://app.dosu.dev/response-feedback/ca29ad17-93a3-4a12-85c9-70e99ca8c9b3?feedback_type=other)</sup> [![Chat with Dosu](https://dosu.dev/dosu-chat-badge.svg)](https://app.dosu.dev/6088a4f4-073a-43f3-b3d4-4f06fb0c05a0/ask?utm_source=github)&nbsp;[![Open in Cursor](https://dosu.dev/dosu-in-cursor.svg)](https://cursor.com/link/prompt?text=The%20most%20likely%20causes%20for%20workflow-run%20outputs%20not%20appearing%20in%20the%20conversation%20component%20are%3A%0A%0A1.%20The%20backend%20only%20saves%20and%20exposes%20answers%20to%20the%20conversation%20UI%20if%20the%20workflow%20run%20completes%20normally%20and%20fires%20the%20%60QueueAdvancedChatMessageEndEvent%60.%20If%20the%20workflow%20is%20stopped%20early%20%28by%20moderation%2C%20annotation%20reply%2C%20manual%20stop%2C%20or%20error%29%2C%20this%20event%20may%20not%20fire%2C%20and%20the%20answer%20won%27t%20be%20saved%20or%20shown%20in%20the%20conversation%E2%80%94even%20if%20the%20workflow-run%20itself%20produced%20output.%20Recent%20fixes%20ensure%20that%20all%20stop%20endpoints%20properly%20record%20workflow%20status%20as%20STOPPED%2C%20but%20if%20your%20deployment%20is%20missing%20these%2C%20workflows%20may%20get%20stuck%20in%20a%20running%20state%20and%20not%20finalize%20the%20answer%20field%20as%20expected%20%5B%28details%29%5D%28https%3A//github.com/langgenius/dify/pull/27803%29.%0A%0A2.%20There%20is%20a%20known%20race%20condition%20in%20the%20output%20moderation%20pipeline%3A%20if%20moderation%20threads%20finish%20after%20the%20main%20event%20loop%20ends%2C%20moderation%20events%20%28and%20possible%20answer%20updates%29%20can%20be%20missed%2C%20leading%20to%20empty%20or%20missing%20responses%20in%20the%20UI.%20The%20suggested%20fix%20is%20to%20delay%20queue%20termination%20until%20all%20moderation%20threads%20complete%20%5B%28details%29%5D%28https%3A//github.com/langgenius/dify/issues/29033%29.%0A%0A3.%20In%20self-hosted%20Docker%20environments%2C%20the%20default%20Nginx%20configuration%20often%20blocks%20Server-Sent%20Events%20%28SSE%29%2C%20causing%20streaming%20output%20and%20conversation%20display%20issues.%20To%20fix%20this%2C%20update%20your%20Nginx%20config%20for%20relevant%20endpoints%20%28like%20%60/api%60%20and%20%60/console/api%60%29%20to%20use%20HTTP/1.1%2C%20disable%20proxy%20buffering%2C%20and%20set%20high%20proxy_read_timeout%20and%20proxy_send_timeout%20values%20%28e.g.%2C%203600s%29.%20Avoid%20overlapping%20location%20blocks%20for%20the%20same%20endpoint%2C%20and%20make%20sure%20HTTP/2%20is%20not%20enabled%20for%20SSE%20endpoints%20%5B%28details%29%5D%28https%3A//github.com/langgenius/dify/issues/19324%29%20%5B%28nginx%20config%20advice%29%5D%28https%3A//github.com/langgenius/dify/issues/24848%29.%0A%0ATo%20troubleshoot%3A%20%0A-%20Test%20with%20different%20input%20types%20%28plain%20text%2C%20various%20file%20uploads%29%20to%20see%20if%20the%20issue%20is%20input-specific.%0A-%20Check%20backend%20logs%20for%20abnormal%20workflow%20stops%2C%20moderation%20events%2C%20or%20errors%20that%20could%20prevent%20the%20answer%20from%20being%20saved.%0A-%20Review%20your%20Nginx%20configuration%20as%20described%20above.%0A-%20Make%20sure%20your%20deployment%20includes%20all%20recent%20fixes%20for%20workflow%20stop%20handling%20and%20moderation%20race%20conditions.%0A%0AIf%20the%20problem%20persists%2C%20capture%20a%20full%20workflow%20run%20%28backend%20logs%2C%20API%20responses%2C%20proxy%20logs%29%20for%20further%20analysis.)&nbsp;[![Join Discord](https://img.shields.io/badge/join-5865F2?logo=discord&logoColor=white&label=)](https://go.dosu.dev/discord-bot)&nbsp;[![Share on X](https://img.shields.io/badge/X-share-black)](https://twitter.com/intent/tweet?text=%40dosu_ai%20helped%20me%20solve%20this%20issue!&url=https%3A//github.com/langgenius/dify/issues/29976)
Author
Owner

@arcsun commented on GitHub (Dec 23, 2025):

@dosu When using the agent node with memory enabled, such blank outputs in the conversation will cause errors in all subsequent questions

Failed to transform agent message: req_id: 70ec83b7e8 PluginInvokeError: {"args":{},"error_type":"Exception","message":"read llm model failed: request failed: req_id: 323825f842 PluginInvokeError: {\"args\":{\"description\":\"[models] Error: API request failed with status code 400: {\\\"error\\\":{\\\"message\\\":\\\"litellm.BadRequestError: BedrockException - {\\\\\\\"message\\\\\\\":\\\\\\\"The text field in the ContentBlock object at messages.1.content.0 is blank. Add text to the text field, and try again.\\\\\\\"}. Received Model Group=claude-sonnet-4-5-20250929\\\\nAvailable Model Group Fallbacks=None\\\",\\\"type\\\":null,\\\"param\\\":null,\\\"code\\\":\\\"400\\\"}}\"},\"error_type\":\"InvokeError\",\"message\":\"[models] Error: API request failed with status code 400: {\\\"error\\\":{\\\"message\\\":\\\"litellm.BadRequestError: BedrockException - {\\\\\\\"message\\\\\\\":\\\\\\\"The text field in the ContentBlock object at messages.1.content.0 is blank. Add text to the text field, and try again.\\\\\\\"}. Received Model Group=claude-sonnet-4-5-20250929\\\\nAvailable Model Group Fallbacks=None\\\",\\\"type\\\":null,\\\"param\\\":null,\\\"code\\\":\\\"400\\\"}}\"}"}
@arcsun commented on GitHub (Dec 23, 2025): [@dosu](https://go.dosu.dev/dosubot) When using the agent node with memory enabled, such blank outputs in the conversation will cause errors in all subsequent questions ``` Failed to transform agent message: req_id: 70ec83b7e8 PluginInvokeError: {"args":{},"error_type":"Exception","message":"read llm model failed: request failed: req_id: 323825f842 PluginInvokeError: {\"args\":{\"description\":\"[models] Error: API request failed with status code 400: {\\\"error\\\":{\\\"message\\\":\\\"litellm.BadRequestError: BedrockException - {\\\\\\\"message\\\\\\\":\\\\\\\"The text field in the ContentBlock object at messages.1.content.0 is blank. Add text to the text field, and try again.\\\\\\\"}. Received Model Group=claude-sonnet-4-5-20250929\\\\nAvailable Model Group Fallbacks=None\\\",\\\"type\\\":null,\\\"param\\\":null,\\\"code\\\":\\\"400\\\"}}\"},\"error_type\":\"InvokeError\",\"message\":\"[models] Error: API request failed with status code 400: {\\\"error\\\":{\\\"message\\\":\\\"litellm.BadRequestError: BedrockException - {\\\\\\\"message\\\\\\\":\\\\\\\"The text field in the ContentBlock object at messages.1.content.0 is blank. Add text to the text field, and try again.\\\\\\\"}. Received Model Group=claude-sonnet-4-5-20250929\\\\nAvailable Model Group Fallbacks=None\\\",\\\"type\\\":null,\\\"param\\\":null,\\\"code\\\":\\\"400\\\"}}\"}"} ```
Author
Owner

@dosubot[bot] commented on GitHub (Jan 24, 2026):

Hi, @zscAZA. I'm Dosu, and I'm helping the Dify team manage their backlog and am marking this issue as stale.

Issue Summary:

  • You reported that in Dify v1.11.1 self-hosted via Docker, workflow runs produce output but the conversation UI does not display responses.
  • Possible causes discussed include incomplete workflow stops, race conditions in output moderation, and Nginx config issues blocking Server-Sent Events.
  • Additional input noted that using the agent node with memory enabled can cause blank outputs, leading to errors on subsequent queries.
  • The issue remains unresolved with no recent updates or confirmed fixes.

Next Steps:

  • Please let me know if this issue is still relevant with the latest version of Dify by commenting here to keep the discussion open.
  • Otherwise, I will automatically close this issue in 15 days.

Thank you for your understanding and contribution!

@dosubot[bot] commented on GitHub (Jan 24, 2026): Hi, @zscAZA. I'm [Dosu](https://dosu.dev), and I'm helping the Dify team manage their backlog and am marking this issue as stale. **Issue Summary:** - You reported that in Dify v1.11.1 self-hosted via Docker, workflow runs produce output but the conversation UI does not display responses. - Possible causes discussed include incomplete workflow stops, race conditions in output moderation, and Nginx config issues blocking Server-Sent Events. - Additional input noted that using the agent node with memory enabled can cause blank outputs, leading to errors on subsequent queries. - The issue remains unresolved with no recent updates or confirmed fixes. **Next Steps:** - Please let me know if this issue is still relevant with the latest version of Dify by commenting here to keep the discussion open. - Otherwise, I will automatically close this issue in 15 days. Thank you for your understanding and contribution!
Author
Owner

@crazywoola commented on GitHub (Feb 3, 2026):

Hi @zscAZA, thanks for opening this issue.

Why this is being closed

This report does not yet meet the required issue standard for langgenius/dify.

  • Steps to reproduce should include clear text descriptions, not only an image.

Relevant guidelines

Next steps

Please open a new issue in English and include all required details from the bug template/contributing guide.

Thanks for understanding and for your contribution.

@crazywoola commented on GitHub (Feb 3, 2026): Hi @zscAZA, thanks for opening this issue. ### Why this is being closed This report does not yet meet the required issue standard for `langgenius/dify`. - Steps to reproduce should include clear text descriptions, not only an image. ### Relevant guidelines - Bug report template: https://github.com/langgenius/dify/blob/3aecceff27c6b712628ad463c6e6ac15b8527ebe/.github/ISSUE_TEMPLATE/bug_report.yml - Code of Conduct / Language Policy: https://github.com/langgenius/dify/blob/4c1ad40f8e8a6ee58a958330558f2178b7e47fa7/.github/CODE_OF_CONDUCT.md - Contributing guide: https://github.com/langgenius/dify/blob/25ac69afc5ac9324079be5f0d02b2a2b03dcc784/CONTRIBUTING.md ### Next steps Please open a new issue in English and include all required details from the bug template/contributing guide. Thanks for understanding and for your contribution.
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: langgenius/dify#21200