Skip to main content

Traefik Timeout in GitLab CI Pipeline

· One min read
Stephan Hochdörfer
Stephan Hochdörfer
Head of IT Business Operations

Recently, failures occurred randomly in our GitLab CI pipelines when pushing docker images to our internal Docker registry.

In the GitLab CI output, all I saw was the following error:

error parsing HTTP 499 response body: invalid character 'C' looking for 
beginning of value: "Client Closed Request"

At first, I was clueless because we hadn't made any major changes in our internal infrastructure, and since the failure happened randomly, it was not easy to identify the source of the problem.

But since I was dealing with a similar problem earlier, I concluded it could be an issue with Traefik. And since I found a similar problem description in one of the GitHub issues, it looked like I was on the right track.

It seems in a Traefik update, the team decided to change the default values for the request timeouts from "no timeout" to 60 seconds.

To solve the issue, I reconfigured the timeouts for our "websecure" entrypoint like this:

[entryPoints.websecure.transport.respondingTimeouts]
readTimeout = 0
writeTimeout = 0
idleTimeout = 0