GitLab Next

Related to #2737

When parsing large codeclimate.json files, the widget experiences significant slow downs.

Multiple issues seem to contribute to this:

• Duplication within the json file created by the codeclimate engine
  • content -> body is duplicated across entries
  • entries are duplicated
• FE code requests the file twice, one for head, another for base, then compares
• Comparing the files in FE takes a long time
• Large files are not compressed, user suggested gzip and more aggressive compression

Todos:

• Collect sample codeclimate.json files of various sizes
  • .1 MB
    • .16 MB: 2,229ms
  • .5 MB
  • 1 MB
  • 2 MB
  • 5 MB
    • 6.75MB: 594ms
  • 10 MB
    • 15MB: 587ms, 900ms
  • 25 MB
• @zeffmorgan : Benchmark local performance with various sized codeclimate.json files
• @drewcimino : Explore Rails reported performance via Kibana
• Report back to #2737 (comment 258922597)