-
Notifications
You must be signed in to change notification settings - Fork 135
Description
Bug Description
Hello, I am encountering sporadic timeouts with gitlab-mcp.
This issue started immediately after upgrading the gitlab-mcp
Docker image from version 1.0.65 to 1.0.76 I also tested version 2.0.2 and experienced the same problem.
The application experiences sporadic timeouts when executing tools. The error returned is MCP error -32001: Request timed out
.
Error Logs:
Failed to execute tool "get_commit"
MCP error -32001: Request timed out
Failed to execute tool "list_commits"
MCP error -32001: Request timed out
Downgrading the image back to 1.0.65 resolves the issue completely, and the application functions as expected. My setup is running within Docker with SSE endpoint.
Steps to Reproduce
- Configure and run the
gitlab-mcp
project using a Docker image with a version greater than1.0.65
(e.g.,1.0.76
or2.0.2
). - Use the configuration provided below, with
SSE
enabled. - Make calls to the service that execute tools like
list_commits
orget_commit
. - Observe the application over time. Sporadically, these tool executions will fail with a timeout error.
Configuration
Here is the relevant environment configuration for the Docker container:
- name: LOG_LEVEL
value: debug
- name: GITLAB_API_URL
value: https://gitlab.com/api/v4
- name: GITLAB_READ_ONLY_MODE
value: "true"
- name: USE_PIPELINE
value: "true"
- name: USE_MILESTONE
value: "true"
- name: USE_GITLAB_WIKI
value: "false"
- name: SSE
value: "true"
- name: HOST
value: 0.0.0.0
- name: PORT
value: "9000"
Troubleshooting Steps Taken
-
Enabled LOG_LEVEL: debug, but no specific errors or indicative logs were found related to the timeouts.
-
Confirmed that downgrading to 1.0.65 is a reliable workaround.
Any guidance on how to further debug this would be greatly appreciated. Thank you.