Handling rate limited requests
A GitLab instance may be behind a reverse proxy that has rate-limiting on API requests to prevent abuse. GitLab Runner sends multiple requests to the API and could go over these rate limits. As a result, GitLab Runner handles rate limited scenarios with the following logic:
- A response code of 429 - TooManyRequests is received.
- The response headers are checked for a
RateLimit-ResetTime
header. TheRateLimit-ResetTime
header should have a value which is a valid HTTP Date (RFC1123), likeWed, 21 Oct 2015 07:28:00 GMT
.- If the header is present and has a valid value the Runner waits until the specified time and issues another request.
- If the header is present, but isn’t a valid date, a fallback of 1 minute is used.
- If the header is not present, no additional actions are taken, the response error is returned.
- The process above is repeated 5 times, then a
gave up due to rate limit
error is returned.
Note:
The header
RateLimit-ResetTime
is case insensitive since all header keys are run
through the http.CanonicalHeaderKey
function.
Help and feedback
If there's something you don't like about this feature
To propose functionality that GitLab does not yet offer
To further help GitLab in shaping new features
If you didn't find what you were looking for
If you want help with something very specific to your use case, and can use some community support
POST ON GITLAB FORUM
If you have problems setting up or using this feature (depending on your GitLab subscription)
REQUEST SUPPORT
To view all GitLab tiers and features or to upgrade
If you want to try all features available in GitLab.com
If you want to try all features available in GitLab self-managed
If you spot an error or a need for improvement and would like to fix it yourself in a merge request
EDIT THIS PAGE
If you would like to suggest an improvement to this doc