Performance Regression or Improvement: sideinpts_python_batch_1gb_1kb_10workers_1000window_1key_percent_dict:runtime

[github-actions]

Performance change found in the
test: sideinpts_python_batch_1gb_1kb_10workers_1000window_1key_percent_dict for the metric: runtime.

For more information on how to triage the alerts, please look at
Triage performance alert issues section of the README.

Test description: python | Side Input | 1 GB dictionary, 1% of keys, 1000 fixed windows
Test Link -

beam/.test-infra/jenkins/job_LoadTests_SideInput_Python.groovy

Line 120 in 5e38dec

name: '1gb-1kb-10workers-1000window-1key-percent-dict',

Test Dashboard - http://metrics.beam.apache.org/d/-E9aGlFGk/side-input-load-tests?orgId=1&from=now-90d&to=now&viewPanel=8

timestamp: Sat Feb  7 18:50:29 2026, metric_value: 945.00
timestamp: Fri Feb  6 19:30:33 2026, metric_value: 1062.00
timestamp: Thu Feb  5 19:18:17 2026, metric_value: 1057.00 <---- Anomaly
timestamp: Wed Feb  4 19:30:40 2026, metric_value: 848.00
timestamp: Tue Feb  3 19:09:05 2026, metric_value: 856.00
timestamp: Mon Feb  2 19:20:47 2026, metric_value: 864.00
timestamp: Sun Feb  1 19:07:53 2026, metric_value: 764.00
timestamp: Sat Jan 31 19:04:48 2026, metric_value: 857.00
timestamp: Fri Jan 30 19:19:06 2026, metric_value: 877.00
timestamp: Thu Jan 29 19:09:06 2026, metric_value: 989.00
timestamp: Wed Jan 28 18:58:15 2026, metric_value: 855.00
timestamp: Tue Jan 27 19:14:08 2026, metric_value: 709.00
timestamp: Mon Jan 26 19:03:11 2026, metric_value: 745.00

[alisha-1000]

I can help triage this alert.

From the dashboard, there is a noticeable spike around Feb 5–6, but values appear to fluctuate in the surrounding runs. I’ll check recent Python SideInput or infra-related changes to see if there is a clear regression signal or if this could be run-to-run variance.

Will update with findings.