fix(ci): increase operator build free-disk-space to 40GB#19396
Conversation
|
Skipping CI for Draft Pull Request. |
There was a problem hiding this comment.
Hey - I've left some high level feedback:
- Since the new
df -h /steps are being used for diagnostics, consider addingif: always()to the ones after long-running build/test steps so they still run and collect disk data even when a preceding step fails.
Prompt for AI Agents
Please address the comments from this code review:
## Overall Comments
- Since the new `df -h /` steps are being used for diagnostics, consider adding `if: always()` to the ones after long-running build/test steps so they still run and collect disk data even when a preceding step fails.Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.
Azure-based GitHub runners can start with only 77GB root disk, leaving ~19GB free. The preamble cleanup stops at the threshold, so 30GB left the operator Docker build (go mod download + UBI base image + source copy + compile) without enough headroom. Bumping to 40GB is safe per preamble docs and consistent with other jobs (emailsender uses 40GB, scanner uses 50GB). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
Images are ready for the commit at 9ec0403. To use with deploy scripts, first |
7f07721 to
9ec0403
Compare
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## master #19396 +/- ##
=======================================
Coverage 49.68% 49.68%
=======================================
Files 2700 2700
Lines 203278 203278
=======================================
Hits 100999 100999
- Misses 94753 94754 +1
+ Partials 7526 7525 -1
Flags with carried forward coverage won't be shown. Click here to find out more. ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
|
I think I was just hit by this issue https://github.com/stackrox/stackrox/actions/runs/23043750790/job/66927778020?pr=19379 |
|
may be not necessary when we apply #19417 to reduce the disk space used by the operator build 🥳 |
@janisz Can I get your +1 for this, and then #19417 will follow-up to reduce the disk usage? Or do you want to keep the 30gb setting because the changes will keep it under 30GB (and go in soon so it doesn't hit/block other PRs)? |
Description
Our operator build used to fit in 30GB and now it uses 31GB.
Azure-based GitHub Actions runners can start with only 77GB root disk (~19GB free). The job-preamble cleanup stops as soon as it reaches the threshold, so the previous 30GB target left the operator Docker build without enough headroom. The multi-stage build (UBI go-toolset base image + full monorepo
go mod download+ source copy + compile) needs more than 30GB.Bumping to 40GB is safe per the preamble's own docs ("ubuntu-24.04 runner starts with 20GB free and we can delete and reach 40GB") and consistent with other jobs (
emailsenderuses 40GB,scanner-versioned-definitionsuses 50GB).See also #19397 for a background disk usage monitor added to the job-preamble to track disk consumption over time.
Disk usage measurements (RHACS amd64, 145G runner)
Peak build consumption: ~31G. On the failing 77G runner where cleanup only freed 30G, this build would have exhausted disk space. The 40GB threshold provides ~9G of headroom.
User-facing documentation
Testing and quality
Automated testing
How I validated my change
dfinstrumentation (see table above)Generated with Claude Code