Add a workflow to run the versions benchmark on demand - #1093
Merged
Conversation
A manual (workflow_dispatch) workflow that launches the versions benchmark for a given ClickHouse version, a version prefix, or master. The runner only launches: it assumes the same federated IAM role as the other benchmark workflows and calls versions/run-benchmark.sh once per (version, machine type). Each machine downloads the prepared Native files, runs run-version.sh, sends its result to the sink and self-terminates. master needs no script changes: list-versions.sh already maps it to the "local" provider, which installs the development build on the machine. The datasets / tries / timeout inputs fall through to the script's defaults when empty, and the repository and branch default to the ones the workflow was dispatched from, so an unmerged branch can be benchmarked as is. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Adds
.github/workflows/versions-benchmark.yml: a manual (workflow_dispatch) workflow that runs the versions benchmark for a specified ClickHouse version — or formaster.The runner only launches, like the other benchmark workflows: it assumes the same federated
ClickBenchLauncherIAM role and callsversions/run-benchmark.shonce per (version, machine type). Each machine downloads the prepared Native files, runsrun-version.sh, POSTs the result (kind: versions-benchmark) and the log to the sink, and self-terminates.Inputs:
versions(required, defaultmaster) — space-separated; an exact version (26.6.1.1193,1.1.54378), a prefix resolved to the newest match (26.6,24), ormaster.machines(defaultc7a.4xlarge),datasets,tries,timeout— empty falls through to the script's own defaults (all ten datasets, 6 tries, 18000 s).repo/branch— default to the repository and branch the workflow was dispatched from, so an unmerged branch can be benchmarked without passing anything.masterneeded no script changes:list-versions.shalready maps it to thelocalprovider, which installs the development build on the machine withcurl https://clickhouse.com/ | shinstead of using Docker.Unknown versions fail at resolution before any AWS call, and per-launch failures are collected and reported at the end, as in
launch-benchmark/action.yml.Verified locally: the YAML parses, the launch step is syntactically valid, the loop expands correctly over versions × machines with the optional inputs empty (dry-run against a stub
run-benchmark.sh), and a bogus version fails the step without touching EC2. A real launch was not exercised.🤖 Generated with Claude Code