issue: Slow knowledge listing #5496

New Issue

2025-11-11T16:22:39-06:00

GiteaMirror commented

2025-11-11 16:22:39 -06:00

Originally created by @a3957273 on GitHub (Jun 10, 2025).

Check Existing Issues

I have searched the existing issues and discussions.
I am using the latest version of Open WebUI.

Installation Method

Docker

Open WebUI Version

0.6.14

Ollama Version (if applicable)

N/A

Operating System

N/A

Browser (if applicable)

Chrome 139

Confirmation

I have read and followed all instructions in README.md.
I am using the latest version of both Open WebUI and Ollama.
I have included the browser console logs.
I have included the Docker container logs.
I have provided every relevant configuration, setting, and environment variable used in my setup.
I have clearly listed every relevant configuration, custom setting, environment variable, and command-line option that influences my setup (such as Docker Compose overrides, .env values, browser settings, authentication configurations, etc).
I have documented step-by-step reproduction instructions that are precise, sequential, and leave nothing to interpretation. My steps:
Start with the initial platform/version/OS and dependencies used,
Specify exact install/launch/configure commands,
List URLs visited, user input (incl. example values/emails/passwords if needed),
Describe all options and toggles enabled or changed,
Include any files or environmental changes,
Identify the expected and actual result at each stage,
Ensure any reasonably skilled user can follow and hit the same issue.

Expected Behavior

I am able to query /api/v1/knowledge/list and it takes under a second to load.

Actual Behavior

Querying /api/v1/knowledge takes 9.5 seconds to load, and during this time OpenWebUI will not respond to any other API requests (e.g. listing chat history, listing users). The application appears to be entirely frozen for those 9.5 seconds.

Steps to Reproduce

Begin with a clean OpenWebUI deployment that uses a Postgres database that's hosted externally.
Upload 128 knowledge entries.
Upload 15 files to each knowledge entry.
Query /api/v1/knowledge/list and observe that it takes 9.5 seconds to load.

Logs & Screenshots

We do not include browser / Docker logs here due to it being a single API call. The browser does not log anything, as it's getting a JSON endpoint. The Docker logs include only a single 200 response for this endpoint, they do not include details, time taken, etc.

Additional Information

I have tried running queries directly against Postgres and seen response times in in the millisecond range, so do not believe it to be the database.

Originally created by @a3957273 on GitHub (Jun 10, 2025). ### Check Existing Issues - [x] I have searched the existing issues and discussions. - [x] I am using the latest version of Open WebUI. ### Installation Method Docker ### Open WebUI Version 0.6.14 ### Ollama Version (if applicable) N/A ### Operating System N/A ### Browser (if applicable) Chrome 139 ### Confirmation - [x] I have read and followed all instructions in `README.md`. - [x] I am using the latest version of **both** Open WebUI and Ollama. - [x] I have included the browser console logs. - [x] I have included the Docker container logs. - [x] I have **provided every relevant configuration, setting, and environment variable used in my setup.** - [x] I have clearly **listed every relevant configuration, custom setting, environment variable, and command-line option that influences my setup** (such as Docker Compose overrides, .env values, browser settings, authentication configurations, etc). - [x] I have documented **step-by-step reproduction instructions that are precise, sequential, and leave nothing to interpretation**. My steps: - Start with the initial platform/version/OS and dependencies used, - Specify exact install/launch/configure commands, - List URLs visited, user input (incl. example values/emails/passwords if needed), - Describe all options and toggles enabled or changed, - Include any files or environmental changes, - Identify the expected and actual result at each stage, - Ensure any reasonably skilled user can follow and hit the same issue. ### Expected Behavior I am able to query /api/v1/knowledge/list and it takes under a second to load. ### Actual Behavior Querying /api/v1/knowledge takes 9.5 seconds to load, and during this time OpenWebUI will not respond to any other API requests (e.g. listing chat history, listing users). The application appears to be entirely frozen for those 9.5 seconds. ### Steps to Reproduce 1. Begin with a clean OpenWebUI deployment that uses a Postgres database that's hosted externally. 2. Upload 128 knowledge entries. 3. Upload 15 files to each knowledge entry. 4. Query /api/v1/knowledge/list and observe that it takes 9.5 seconds to load. ### Logs & Screenshots We do not include browser / Docker logs here due to it being a single API call. The browser does not log anything, as it's getting a JSON endpoint. The Docker logs include only a single 200 response for this endpoint, they do not include details, time taken, etc. ### Additional Information I have tried running queries directly against Postgres and seen response times in in the millisecond range, so do not believe it to be the database.

GiteaMirror added the bug label 2025-11-11 16:22:39 -06:00

GiteaMirror closed this issue

2025-11-11 16:22:40 -06:00

GiteaMirror commented

2025-11-11 16:22:40 -06:00

@a3957273 commented on GitHub (Jun 10, 2025):

To reproduce this you should add a few hundred milliseconds of latency to connections to the database. Either by having it hosted over a network, or by artificially adding it via iptable configuration, etc.

@a3957273 commented on GitHub (Jun 10, 2025): To reproduce this you should add a few hundred milliseconds of latency to connections to the database. Either by having it hosted over a network, or by artificially adding it via iptable configuration, etc.

GiteaMirror commented

2025-11-11 16:22:40 -06:00

@germanoquanta commented on GitHub (Jun 10, 2025):

I don't know if you've just updated to this version too, I was in 0.6.13 and after the update to v0.6.14 (i used containerrr watchtower), it slowed down the answer speed. Tried with different models too. Im using Windows 11 Desktop Docker, no Postgres external connection, just the Ollama that's installed in my Windows.

@germanoquanta commented on GitHub (Jun 10, 2025): I don't know if you've just updated to this version too, I was in 0.6.13 and after the update to v0.6.14 (i used containerrr watchtower), it slowed down the answer speed. Tried with different models too. Im using Windows 11 Desktop Docker, no Postgres external connection, just the Ollama that's installed in my Windows.

GiteaMirror commented

2025-11-11 16:22:41 -06:00