langgenius/dify
1
0
Fork 0
mirror of https://github.com/langgenius/dify.git synced 2026-07-21 12:05:30 -04:00
Code Issues 775 Packages Projects Releases 10 Wiki Activity

Large-scale workflows often lose node configurations. #21499

New Issue
Open
opened 2026-02-21 20:12:51 -05:00 by yindo · 2 comments
yindo commented 2026-02-21 20:12:51 -05:00
Owner
Copy Link
Copy Source

Originally created by @WilliamChen-luckbob on GitHub (Jan 7, 2026).

Originally assigned to: @Nov1c444 on GitHub.

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

When a workflow is complex and large-scale, featuring a large number of agents, branches, and Python custom scripts, the loss of previously configured content often occurs, and we are unsure how to reproduce this issue.

✔️ Expected Behavior

Every time the Publish button is clicked, the workflow on the current canvas should be saved as my latest configuration.

❌ Actual Behavior

In fact, even so—even after I have tested and validated the workflow—some configurations of my nodes (such as node names, variable names, and even Python scripts) have been reset after publishing and refreshing the page.

you can see in the images, the python script was gone when I open that page again.

Image Image
Originally created by @WilliamChen-luckbob on GitHub (Jan 7, 2026). Originally assigned to: @Nov1c444 on GitHub. ### 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 When a workflow is complex and large-scale, featuring a large number of agents, branches, and Python custom scripts, the loss of previously configured content often occurs, and we are unsure how to reproduce this issue. ### ✔️ Expected Behavior Every time the Publish button is clicked, the workflow on the current canvas should be saved as my latest configuration. ### ❌ Actual Behavior In fact, even so—even after I have tested and validated the workflow—some configurations of my nodes (such as node names, variable names, and even Python scripts) have been reset after publishing and refreshing the page. you can see in the images, the python script was gone when I open that page again. <img width="1311" height="844" alt="Image" src="https://github.com/user-attachments/assets/23aa6822-f449-4a1c-89e7-4237bcb0eeef" /> <img width="1331" height="768" alt="Image" src="https://github.com/user-attachments/assets/e2da137d-2487-488e-b99d-df8ba61b04ae" />
yindo added the 🐞 bugstale labels 2026-02-21 20:12:51 -05:00
yindo commented 2026-02-21 20:12:52 -05:00
Author
Owner
Copy Link
Copy Source

@WilliamChen-luckbob commented on GitHub (Jan 7, 2026):

@dosu
This may be caused by the page being in edit mode for an extended period.

I have been trying to edit and test the workflow since this morning, accumulating a large number of operation traces in the history (possibly hundreds of operations).

This issue happened to occurr repeatedly during the tests this afternoon. When I clicked Preview to start testing the workflow (without refreshing the page), I found that the configurations of some previously set up workflow nodes became corrupted—such as reverting to previous variable names, losing functions entirely, or having prompts reverted.

I have the habit of clicking Publish before each preview test; therefore, I believe this may be related to abnormal page caching.

@WilliamChen-luckbob commented on GitHub (Jan 7, 2026): @dosu This may be caused by the page being in edit mode for an extended period. I have been trying to edit and test the workflow since this morning, accumulating a large number of operation traces in the history (possibly hundreds of operations). This issue happened to occurr repeatedly during the tests this afternoon. When I clicked Preview to start testing the workflow (without refreshing the page), I found that the configurations of some previously set up workflow nodes became corrupted—such as reverting to previous variable names, losing functions entirely, or having prompts reverted. I have the habit of clicking Publish before each preview test; therefore, I believe this may be related to abnormal page caching.
yindo commented 2026-02-21 20:12:52 -05:00
Author
Owner
Copy Link
Copy Source

@dosubot[bot] commented on GitHub (Jan 7, 2026):

Your observations are spot on—long editing sessions and large operation histories can cause workflow state loss or node configuration corruption in Dify, especially in complex workflows. This is a known issue, and recent fixes (merged after v1.11.1) specifically address problems with workflow version synchronization and session storage handling during preview, publishing, and extended editing. Upgrading to a newer version should help prevent these issues [PR #30139, PR #30300].

In the meantime, here are some practical steps to help mitigate or diagnose the problem:

  • Periodically refresh the page during long editing sessions to reset the frontend state.
  • Keep workflows moderate in size if possible, or split very large workflows into smaller ones to reduce UI and state management stress [performance discussion].
  • Check your browser console and backend logs for errors during publish or preview—failed API calls or connection resets can indicate infrastructure or resource issues [infra checklist].
  • Ensure your Docker volumes are correctly mapped and persistent, and that all database migrations have completed successfully. Sometimes changes appear saved but revert after refresh if persistence is misconfigured [backend persistence].
  • Tune environment variables like SESSION_EXPIRE_HOURS, REDIS_CACHE_TTL, and WORKFLOW_NODE_OUTPUT_TTL to extend session and cache retention [session/caching].
  • If you see a node panel disappear, check if it’s just visually collapsed (look for a thin border to drag open) rather than truly lost [UI bug workaround].

If you can share any specific error messages or logs from when the issue occurs, I can help interpret them to further isolate the root cause.

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 (Jan 7, 2026): <!-- Answer --> Your observations are spot on—long editing sessions and large operation histories can cause workflow state loss or node configuration corruption in Dify, especially in complex workflows. This is a known issue, and recent fixes (merged after v1.11.1) specifically address problems with workflow version synchronization and session storage handling during preview, publishing, and extended editing. Upgrading to a newer version should help prevent these issues [[PR #30139](https://github.com/langgenius/dify/pull/30139), [PR #30300](https://github.com/langgenius/dify/pull/30300)]. In the meantime, here are some practical steps to help mitigate or diagnose the problem: - Periodically refresh the page during long editing sessions to reset the frontend state. - Keep workflows moderate in size if possible, or split very large workflows into smaller ones to reduce UI and state management stress [[performance discussion](https://github.com/langgenius/dify/issues/28293)]. - Check your browser console and backend logs for errors during publish or preview—failed API calls or connection resets can indicate infrastructure or resource issues [[infra checklist](https://github.com/langgenius/dify/issues/23381)]. - Ensure your Docker volumes are correctly mapped and persistent, and that all database migrations have completed successfully. Sometimes changes appear saved but revert after refresh if persistence is misconfigured [[backend persistence](https://github.com/langgenius/dify/issues/25629)]. - Tune environment variables like SESSION_EXPIRE_HOURS, REDIS_CACHE_TTL, and WORKFLOW_NODE_OUTPUT_TTL to extend session and cache retention [[session/caching](https://github.com/langgenius/dify/issues/24616)]. - If you see a node panel disappear, check if it’s just visually collapsed (look for a thin border to drag open) rather than truly lost [[UI bug workaround](https://github.com/langgenius/dify/issues/22107)]. If you can share any specific error messages or logs from when the issue occurs, I can help interpret them to further isolate the root cause. <!-- 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/5a6de595-48ad-423a-953c-1b2062a880c7?feedback_type=great_response) | [Irrelevant](https://app.dosu.dev/response-feedback/5a6de595-48ad-423a-953c-1b2062a880c7?feedback_type=irrelevant_answer) | [Incorrect](https://app.dosu.dev/response-feedback/5a6de595-48ad-423a-953c-1b2062a880c7?feedback_type=incorrect_sources) | [Verbose](https://app.dosu.dev/response-feedback/5a6de595-48ad-423a-953c-1b2062a880c7?feedback_type=too_verbose) | [Hallucination](https://app.dosu.dev/response-feedback/5a6de595-48ad-423a-953c-1b2062a880c7?feedback_type=hallucination) | [Report 🐛](https://app.dosu.dev/response-feedback/5a6de595-48ad-423a-953c-1b2062a880c7?feedback_type=bug_report) | [Other](https://app.dosu.dev/response-feedback/5a6de595-48ad-423a-953c-1b2062a880c7?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=Your%20observations%20are%20spot%20on%E2%80%94long%20editing%20sessions%20and%20large%20operation%20histories%20can%20cause%20workflow%20state%20loss%20or%20node%20configuration%20corruption%20in%20Dify%2C%20especially%20in%20complex%20workflows.%20This%20is%20a%20known%20issue%2C%20and%20recent%20fixes%20%28merged%20after%20v1.11.1%29%20specifically%20address%20problems%20with%20workflow%20version%20synchronization%20and%20session%20storage%20handling%20during%20preview%2C%20publishing%2C%20and%20extended%20editing.%20Upgrading%20to%20a%20newer%20version%20should%20help%20prevent%20these%20issues%20%5B%5BPR%20%2330139%5D%28https%3A//github.com/langgenius/dify/pull/30139%29%2C%20%5BPR%20%2330300%5D%28https%3A//github.com/langgenius/dify/pull/30300%29%5D.%0A%0AIn%20the%20meantime%2C%20here%20are%20some%20practical%20steps%20to%20help%20mitigate%20or%20diagnose%20the%20problem%3A%0A%0A-%20Periodically%20refresh%20the%20page%20during%20long%20editing%20sessions%20to%20reset%20the%20frontend%20state.%0A-%20Keep%20workflows%20moderate%20in%20size%20if%20possible%2C%20or%20split%20very%20large%20workflows%20into%20smaller%20ones%20to%20reduce%20UI%20and%20state%20management%20stress%20%5B%5Bperformance%20discussion%5D%28https%3A//github.com/langgenius/dify/issues/28293%29%5D.%0A-%20Check%20your%20browser%20console%20and%20backend%20logs%20for%20errors%20during%20publish%20or%20preview%E2%80%94failed%20API%20calls%20or%20connection%20resets%20can%20indicate%20infrastructure%20or%20resource%20issues%20%5B%5Binfra%20checklist%5D%28https%3A//github.com/langgenius/dify/issues/23381%29%5D.%0A-%20Ensure%20your%20Docker%20volumes%20are%20correctly%20mapped%20and%20persistent%2C%20and%20that%20all%20database%20migrations%20have%20completed%20successfully.%20Sometimes%20changes%20appear%20saved%20but%20revert%20after%20refresh%20if%20persistence%20is%20misconfigured%20%5B%5Bbackend%20persistence%5D%28https%3A//github.com/langgenius/dify/issues/25629%29%5D.%0A-%20Tune%20environment%20variables%20like%20SESSION_EXPIRE_HOURS%2C%20REDIS_CACHE_TTL%2C%20and%20WORKFLOW_NODE_OUTPUT_TTL%20to%20extend%20session%20and%20cache%20retention%20%5B%5Bsession/caching%5D%28https%3A//github.com/langgenius/dify/issues/24616%29%5D.%0A-%20If%20you%20see%20a%20node%20panel%20disappear%2C%20check%20if%20it%E2%80%99s%20just%20visually%20collapsed%20%28look%20for%20a%20thin%20border%20to%20drag%20open%29%20rather%20than%20truly%20lost%20%5B%5BUI%20bug%20workaround%5D%28https%3A//github.com/langgenius/dify/issues/22107%29%5D.%0A%0AIf%20you%20can%20share%20any%20specific%20error%20messages%20or%20logs%20from%20when%20the%20issue%20occurs%2C%20I%20can%20help%20interpret%20them%20to%20further%20isolate%20the%20root%20cause.)&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/30687)
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
Knowledge Pipeline
Proposal
WIP
cloud
dependencies
feat:webapp
good first issue
outdated
performance
pull-request
Mirrored from GitHub Pull Request
 python:uv
refactor
size:L
stale
status: accepting prs
status: blocked
status: in discussion
stt & tts
vulnerablity
web
wont-fix
🌊 feat:workflow
🌚 invalid
🐞 bug
👻 feat:rag
💪 enhancement
📚 documentation
🔨 feat:tools
🙋‍♂️ question
🙏 help wanted
🤔 cant-reproduce
🤖 feat:agent
No Label stale 🐞 bug
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
yindo
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: langgenius/dify#21499
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?