ci/deploy: add phase markers to devsite.sh#36794
Draft
bosconi wants to merge 1 commit into
Draft
Conversation
Wrap each step of the dev-site deploy in `ci_collapsed_heading` so the Buildkite UI shows per-phase sections and durations, and so log scraping can pin down which sub-step is slow. The script currently runs ~45m end to end with only two intermediate log markers (the two "Docs are in" lines from bin/doc), which makes it impossible to attribute the ~14m of growth over the past 60 days to bin/pydoc, the S3 syncs, or rustdoc without instrumenting first. No behavior change — only adds `. misc/shlib/shlib.bash` and eight `ci_collapsed_heading` lines around the existing commands.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Wraps each step of
ci/deploy/devsite.shinci_collapsed_heading(frommisc/shlib/shlib.bash) so the Buildkite UI gets per-phase sections and so log scraping can pin down which sub-step dominates the deploy time.Motivation
The
deploy-devsiteBuildkite step has grown from ~25m → ~48m p50 over the past 60 days (data: 783 builds on main). It runs serially on a single concurrency slot (concurrency_group: deploy/devsite), so on burst days a 5-10 commit run can queue 4-8 hours of downstream work.Today the script logs only three intermediate markers:
Docs are in /mnt/build/doc/index.htmllines frombin/docpdoc failed to precompile the search indexline frombin/pydocThat is enough to identify two growth modes:
bin/doc(public rustdoc) stepped from 3m to 10m between Apr 15 and Apr 30 and has been stable sinceaws s3 syncx 3 +bin/pydocphase has grown linearly 19m to 33m with no markers in betweenBut it is not enough to attribute that ~14m of linear growth to any one of: aws sync of
target-xcompile/doc/(rustdoc HTML count keeps growing), aws sync oftarget/pydoc/, orbin/pydocitself (the failing search-index step suggests Python doc volume is hitting a scaling limit).Change
Sources
misc/shlib/shlib.bashand adds eightci_collapsed_headinglines, one before each existing command. No command changes, no behavior changes.After a week of post-merge data we will know which phase is the linear-creep driver and can decide whether to invest in faster pdoc, S3 sync parallelism, or path-based skipping for non-docs commits.
Test plan
deploybuild on main shows eight collapsed sections with durationsbash -n ci/deploy/devsite.shpasses (verified locally)