after update to v1.2.0, RAG knowledge retrieve step became extremly slow #12852

Closed
opened 2026-02-21 19:09:23 -05:00 by yindo · 3 comments
Owner

Originally created by @boaty on GitHub (Apr 13, 2025).

Self Checks

  • 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 (我已阅读并同意 Language Policy).
  • [FOR CHINESE USERS] 请务必使用英文提交 Issue,否则会被关闭。谢谢!:)
  • Please do not modify this template :) and fill in all the required fields.

Dify version

1.2.0

Cloud or Self Hosted

Self Hosted (Docker)

Steps to reproduce

i have my dify chatflow with RAG retrive using cohere rerank v3.5. Knowledge base is built by local bge-m3 (ollama) everything was fine, quick and accurate until the update to dify v1.2.0 from v1.0.
the knowledge retrieve step alone took on average 5 mins to finish, compare to ~ 15s in previous version. all data and setting are the same.

please fix this problem.

✔️ Expected Behavior

the prevous difi version worked well with speed and accurate

Actual Behavior

i have to wait 5 mins for knowledge retrieve part

Originally created by @boaty on GitHub (Apr 13, 2025). ### Self Checks - [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 (我已阅读并同意 [Language Policy](https://github.com/langgenius/dify/issues/1542)). - [x] [FOR CHINESE USERS] 请务必使用英文提交 Issue,否则会被关闭。谢谢!:) - [x] Please do not modify this template :) and fill in all the required fields. ### Dify version 1.2.0 ### Cloud or Self Hosted Self Hosted (Docker) ### Steps to reproduce i have my dify chatflow with RAG retrive using cohere rerank v3.5. Knowledge base is built by local bge-m3 (ollama) everything was fine, quick and accurate until the update to dify v1.2.0 from v1.0. the knowledge retrieve step alone took on average 5 mins to finish, compare to ~ 15s in previous version. all data and setting are the same. please fix this problem. ### ✔️ Expected Behavior the prevous difi version worked well with speed and accurate ### ❌ Actual Behavior i have to wait 5 mins for knowledge retrieve part
yindo closed this issue 2026-02-21 19:09:23 -05:00
Author
Owner

@dosubot[bot] commented on GitHub (Apr 13, 2025):

Hello @boaty! I'm Dosu, a bot here to help you with bugs, answer questions, and assist with contributing. I'm here while you wait for a human maintainer. Let's tackle this issue together!

The slowdown in the RAG knowledge retrieval step after updating to Dify version 1.2.0 is due to a known bug in that version. This issue has been identified and fixed, and the fixes will be available in the next release. You will need to wait for the next version update to resolve this performance issue [1].

To continue talking to Dosu, mention @dosu.


Help Dosu learn! Give it feedback: Great Response | Irrelevant Answer | Incorrect Sources | Too Verbose | Hallucination | Bug Report | Other

@dosubot[bot] commented on GitHub (Apr 13, 2025): <!-- Greeting --> Hello @boaty! I'm [Dosu](https://dosu.dev), a bot here to help you with bugs, answer questions, and assist with contributing. I'm here while you wait for a human maintainer. Let's tackle this issue together! <!-- Answer --> The slowdown in the RAG knowledge retrieval step after updating to Dify version 1.2.0 is due to a known bug in that version. This issue has been identified and fixed, and the fixes will be available in the next release. You will need to wait for the next version update to resolve this performance issue <sup>[[1]](https://github.com/langgenius/dify/issues/17921)</sup>. <!-- Dosu Comment Footer --> *To continue talking to [Dosu](https://dosu.dev), mention @dosu.* --- <sup>Help Dosu learn! Give it feedback: [Great Response](https://app.dosu.dev/response-feedback/b41ee34b-ea64-4d07-ad31-d97de3579be2?feedback_type=great_response) | [Irrelevant Answer](https://app.dosu.dev/response-feedback/b41ee34b-ea64-4d07-ad31-d97de3579be2?feedback_type=irrelevant_answer) | [Incorrect Sources](https://app.dosu.dev/response-feedback/b41ee34b-ea64-4d07-ad31-d97de3579be2?feedback_type=incorrect_sources) | [Too Verbose](https://app.dosu.dev/response-feedback/b41ee34b-ea64-4d07-ad31-d97de3579be2?feedback_type=too_verbose) | [Hallucination](https://app.dosu.dev/response-feedback/b41ee34b-ea64-4d07-ad31-d97de3579be2?feedback_type=hallucination) | [Bug Report](https://app.dosu.dev/response-feedback/b41ee34b-ea64-4d07-ad31-d97de3579be2?feedback_type=bug_report) | [Other](https://app.dosu.dev/response-feedback/b41ee34b-ea64-4d07-ad31-d97de3579be2?feedback_type=other)</sup>
Author
Owner

@crazywoola commented on GitHub (Apr 13, 2025):

Already fixed in

  • fix segment query index not effect (#17704)
  • revert batch query (#17707)
    Can you provide how big the embeddings table is?
@crazywoola commented on GitHub (Apr 13, 2025): Already fixed in - fix segment query index not effect (#17704) - revert batch query (#17707) Can you provide how big the embeddings table is?
Author
Owner

@lywy233 commented on GitHub (Jun 20, 2025):

Same problem, the same test takes 0.3s at 0.15 and 1.4s for 1.4.2

@lywy233 commented on GitHub (Jun 20, 2025): Same problem, the same test takes 0.3s at 0.15 and 1.4s for 1.4.2
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: langgenius/dify#12852