Raw Report

Every measurement Locust recorded for the Registry Staff API performance tests, verbatim, as a downloadable file.

Every measurement Locust actually recorded, verbatim: no curation to a headline endpoint set, no SLO comparison, no PASS/FAIL. That interpretation lives in the final report, built from this data.

registry-staff-api-raw-report.zip

archive · 360KB

Open

The full raw report (Markdown, 7 MB uncompressed, 91 tables), zipped.

The raw report is published as a download, not as a page: at this size GitBook cannot render it, and a page that fails to parse stops the documentation from syncing. It is generated by performance-testing/scripts/create_raw_report.py in the Farmer Registry repository from the Locust results (locust/api/results/staff-api/); re-run it after a new test run rather than editing the report by hand.

What it contains

One table per ingress, volume tier, pod scale and test step (and per scenario for step 1), with the columns from Locust's own statistics: request and failure counts, response times (median, average, min, max and percentiles p50 to p100), content size, and requests and failures per second.

• End-to-End

  • Volume-Tier: primary

  • Volume-Tier: smoke

  • Volume-Tier: stretch

• In-Cluster

  • Volume-Tier: primary

  • Volume-Tier: stretch