Skip to content

Update integration tests and retain reproducible performance evidence - #15

Merged
liran merged 4 commits into
mainfrom
randy/integration-performance
Oct 10, 2026
Merged

liran merged 4 commits into
mainfrom
randy/integration-performance

Conversation

@liran

@liran liran commented Oct 10, 2026 •

Copy link
Copy Markdown
Contributor

Integration fixtures used removed server settings and metrics, while benchmark artifacts could lose their connection to the measured source and binary. Update fixtures for the current public SDK/protocol, verify bounded result streaming and drained accounting, and retain frozen source, commands and owned-fixture evidence. Bind every report to its actual binary, source and dependency pins, reject changed binaries and incomplete physical counters, and pin the server to the optimized Weir commit. Search measurements default to 20-second warmup and measurement; server batching remains 32 records.

Validated with race tests, default and integration-tagged vet, offline dependency checks, 19 Python tests in normal and optimized modes, and full MongoDB/Elasticsearch integration against the measured server source. Historical summaries reproduce the original throughput and latency values, retain negative results, and distinguish database drift. Independent reviews cover the server integration contracts and benchmark evidence validation.

@liran liran self-assigned this Oct 10, 2026
@liran
liran merged commit 0a4636f into main Oct 10, 2026
5 checks passed
@liran
liran deleted the randy/integration-performance branch October 10, 2026 15:37
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant