DO NOT MERGE: bit_pr load test on large (4 vCPU / 8GB) - #10604
Closed
luvkapur wants to merge 3 commits into
Closed
Conversation
…r sizing) Insights (30d, all branches): ~4.8M credits/month (~$2.9k) — e2e_test 3.41M (71%), bit_pr 800k (17%), e2e_test_bbit 200k, bit_merge 153k. - stop persisting .pnpm-store: nothing downstream reads it; it inflated the workspace every node of every job attaches (~2.7min x 40 e2e nodes/run) - e2e_test off master pushes: gates nothing there (bit_merge and the merge queue don't depend on it); a daily canary now runs in nightly instead (~440k credits/month -> ~185k) - e2e_test parallelism 40 -> 30: per-node fixed setup is the only cost that scales with node count (suite time is constant); ~4min wall-clock cost - bit_pr 2xlarge -> xlarge: build is CPU-bound on a single env, 16 cores were idle at 80 credits/min; heap capped at 12GB to fit 16GB RAM (~400k/month if it holds — revert to 2xlarge on 'Killed') - e2e_test_bbit halts on quiet days: no new @teambit/bit version in 24h means the previous nightly already bundled+tested the same thing (32 nodes x 30min) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…npm call The gate only ever saves work, so a metadata hiccup (npm outage, missing time entry, non-numeric output) must run the tests, not fail the job. Also one npm invocation instead of two — it runs on every parallel node. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Same 98-component cascade as the xlarge test (comment touch on teambit.component/component), now on large (4 vCPU / 8GB / 20 cr/min, heap 6144). xlarge already passed this in 25 min / ~1,010 credits; this probes the next rung down. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Member
Author
|
Verdict: large (8GB) cannot run the heavy-cascade class — cgroup OOM ('Received "killed" signal') during GenerateEnvTemplate's webpack bundling at minute 16 of the 98-component cascade, after snap/compile/test/BundleUI/ExtractSchema all passed. Wasted-run cost of the probe: ~380 credits. Empirical tier map so far:
Remote lane teambit.bit/ci-pr-loadtest-large can be janitored. |
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.
Next rung down from the xlarge validation (#10602: 98-component cascade, 25 min, ~1,010 credits, no OOM). This branch runs the same cascade with
resource_class: large+ 6GB heap. If it survives with acceptable wall time, dynamic sizing tiers can start at large; if it OOMs or crawls, xlarge is the floor for heavy PRs. Will be closed and branch deleted once the verdict is recorded.🤖 Generated with Claude Code