Skip to content

Artifacts listed in usage quotas don't align with data returned from API

Problem description

We have a project (vyne/vyne - project id 3491723) with very high artifacts listed in the quota page:

image

Analysis

However, using the storage analyzer, I get the following output:

➜ docker run -ti -v "`pwd`/reports:/tmp" \
 -e "GL_TOKEN=xxxxxx" \
 -e "GL_PROJECT_ID=3491723" \
 -e "GL_OUTPUT_FILE=/tmp/gitlab_storage_analyzer_report.md" \
 -e "GL_THRESHOLD_AGE_SEC=86000" \
  -e "GL_VERBOSE=1" \
 registry.gitlab.com/gitlab-de/gitlab-storage-analyzer \
 python /app/gitlab_storage_analyzer.py

Results Summary

Project Statistics

Project Statistics

Project Name Project ID Storage (MB) Repository (MB) Job Artifacts (MB) Packages (MB) Wiki (MB) Snippets (MB) Uploads (MB)
vyne/vyne (Usage quota) 3491723 24176.3450 128.8000 24006.2151 0.0000 0.2100 0.2800 40.8399
Description Size (MB)
Job artifacts size 756.4468 MB
Age threshold (86000.0 sec) Cleanup Potential 0.0000 MB

This lists the job artifacts as both 24GB AND 756MB.

Update @dnsmichi: This has been addressed to be a bug in the script.

Summary

There appear to be job artifacts which are not listed, and therefore cannot be deleted.

This is very concerning, as given the upcoming changes to charges for quota, I'm significantly over the 5GB storage limit, but can't do anything to manage it.

Potential problems

Edited by Michael Friedrich