Performance testing

Release Benchmark

Compare two JMeter runs and see what changed before you ship. Find slower samplers, rising error rates, and gaps in test coverage.

Files stay on your deviceBrowser-only analysis · No account required

Comparison report

Threshold check

Overall error rateBaseline → Candidate

Overall samples/sBaseline → Candidate

Run window

Show Find a sampler

Values are baseline → candidate. Scroll the table for all metrics.

Baseline and candidate metrics by sampler label
Sampler	Status	p95 (ms)	Error rate	Samples/s	Samples

Use comparable runs: Keep the test plan, workload, environment, and ramp-up the same. Compare equivalent sets of samplers. Overall metrics include every row, including transaction controllers if exported.
Use raw sample results: Save comma-delimited results with a header, using epoch milliseconds and start timestamps. Required fields are timeStamp, elapsed, label, and success.
Read the threshold check: p95 uses nearest rank and includes failed samples. Error deltas use percentage points. New, missing, and low-volume samplers need review. Samples/s uses the full run window and is informational.