Artifacts Usage Report — 2026-05-31 #36080
Closed
Replies: 1 comment
-
|
This discussion was automatically closed because it expired on 2026-06-01T07:03:11.677Z.
|
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
Overview
This report analyzes GitHub Actions artifact usage across 68 workflows in
github/gh-awbased on recent runs (last ~5–10 runs per workflow). In total, 1,813 artifacts were found consuming ~18.3 GB of storage. The bulk of storage is dominated by a handful of high-frequency agentic workflows that upload large context/log bundles per run.Summary Stats
Artifact Usage by Workflow
View full table (68 workflows, sorted by total size)
🔍 Insights
Storage Hogs (Top 5 — account for ~60% of total storage)
Most Active (by artifact count)
Stale Workflows (last run > 6 months ago)
These workflows are still marked active but haven't run recently — consider archiving or disabling:
Test Claude(last run: 2025-09-14).github/workflows/test-proxy.lock.yml(last run: 2025-08-29)Test Copilot GitHub Integration(last run: 2025-09-30)copilot only(last run: 2025-11-26).github/workflows/daily-perf-improver.lock.yml(last run: 2025-11-14).github/workflows/daily-test-improver.lock.yml(last run: 2025-11-13)💡 Recommendations
Reduce retention on large daily workflows: Workflows like Daily Issues Report Generator, Daily News, Copilot Session Insights generate 30–72 MB artifacts per run. Set
retention-days: 7(or even 3) instead of the default 90 days — this could save 10+ GB over time.Compress artifacts before upload: NLP/analysis workflows uploading multi-MB JSON/log bundles should use
gziportar.gzbefore callingactions/upload-artifact. Typical compression ratios of 5–10x are achievable for JSON.Disable or archive stale workflows: 6 workflows haven't run since Aug–Nov 2025 — they accumulate metadata and confuse the workflow list without value.
Doc Build - Deploy outlier: 3 artifacts averaging 244 MB each. If these are full site builds, consider using GitHub Pages deployment artifacts with
actions/upload-pages-artifactwhich is optimized for static sites.Smoke tests cleanup: Smoke Copilot (780 MB over 43 runs) and Smoke Codex (488 MB over 25 runs) average 18–20 MB per run. Consider whether full context logs need to be uploaded for every smoke run, or only on failure.
References:
Beta Was this translation helpful? Give feedback.
All reactions