Resolving an HTTP 500 error in Managed Service for GitLab
Issue description
When trying to access a Managed Service for GitLab instance, you see an HTTP 500 or HTTP 502 error message.
Solution
This issue usually occurs when a Managed Service for GitLab instance runs out of free disk space. Container images and build artifacts often take up a significant amount of disk space.
To track the amount of free disk space on your Managed Service for GitLab instance and avoid running out of it moving forward, you may want to set up monitoring for your Managed Service for GitLab instance by following this guide (please note that this feature is still at the Preview stage).
Note
When creating a Managed Service for GitLab instance, Delete untagged images during maintenance is enabled by default. This option enables periodic deletion of images not referenced by build manifests.
To resolve this issue, you need to run the disk partition cleanup as per this guide.
Alternatively, you can increase the Managed Service for GitLab instance’s disk size.
If this does not help, create a support ticket.
When creating a ticket, specify the ID of the problematic Managed Service for GitLab instance and supply the screen capture with error details.