Measure complete Argon workflows with reproducible source provenance - #1
Merged
Merged
Conversation
jakezwang
marked this pull request as ready for review
September 7, 2026 19:10
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Metadata-fork latency does not describe the cost of a usable native MongoDB sandbox. This change measures those boundaries separately: physical sandbox/capture readiness, first native query, acknowledged-write visibility, reads before and after snapshots, and storage after divergence. Each run retains raw samples, percentile summaries and frozen engine/suite source provenance. Historical results keep their original scope.
Requires the companion Argon 2.1.0 candidate commit
ba6e06a9a3b31124d6c37475b5667dd70ab42379, includingStartCapture,SyncBranchandWaitAuto, plus Go 1.26.6 and an isolated MongoDB replica set. Usescripts/run.py --engine <checkout> --ref <commit>; its temporary module replacement selects the reviewed engine whilego.modretains the historical published baseline. PR CI now fetches that exact companion engine commit explicitly. Fixture teardown drains automatic snapshots before deleting its databases.Recorded validation: all eight 1k/10k-document × 1/4-worker × 1/4-ancestry cells completed against the exact core commit and suite commit
a05950938a8eae5ba8cd089a02d9b7207a641945, with 800 metadata-fork samples, 160 sandbox/capture workflows, 20 reads per phase per cell and verified divergence counts. The reports include raw data and source manifests. These are local Go 1.26.6 / MongoDB 7.0.14 measurements; small-sample p99 and allocated storage are scoped accordingly. The expanded scale matrix remains unmeasured; the container recipe is exercised by the focused PR CI smoke below. This PR does not publish a package or claim production performance limits.PR CI now builds the Go 1.26.6 Docker recipe, starts the MongoDB replica set, runs the runner's Go unit tests, and executes one 100-document / one-worker / one-level workflow cell. It verifies the exact engine and suite refs, actual Go version, and six captured divergence updates; no latency or throughput SLA is asserted. Raw samples, provenance, frozen source archives and logs are retained as an Actions artifact for 14 days. The Docker image uses Debian trixie to provide the Python archive-filter API used by the runner, and the CI Bash configuration propagates smoke failures through the log pipeline.
Remote validation passed on head
80d9e9754c63c3f71d6298dd606ce022767e5c5d: Benchmark CI run 34149802845. The previously recorded eight-cell matrix and its measured source provenance are unchanged.