Page MenuHomePhabricator

Investigate the inconsistent load test results (locust) for revertrisk
Closed, ResolvedPublic2 Estimated Story Points

Description

The current load test results for RevertRisk are not reproducible. We suspect the test was run using the API Gateway. To continue using locust as our load testing tool, we need to investigate this issue and correct the test results.

@isarantopoulos do you remember the config values in locust.conf when you ran the revertrisk tests? I can't reproduce the result in revertrisk_stats.csv. I haven't deployed RRLA to staging yet, so it's the same model you tested.

With the current locust.conf, I got the result P58995, which has fewer requests.

I also attempted to use the previous locust.conf, but the result P58996 was still different, with a significantly higher number of requests.

Event Timeline

isarantopoulos set the point value for this task to 2.Apr 9 2024, 2:44 PM
isarantopoulos moved this task from Unsorted to Ready To Go on the Machine-Learning-Team board.

Change #1025805 had a related patch set uploaded (by Ilias Sarantopoulos; author: Ilias Sarantopoulos):

[machinelearning/liftwing/inference-services@main] revertrisk: update locust results

https://gerrit.wikimedia.org/r/1025805

isarantopoulos removed isarantopoulos as the assignee of this task.
isarantopoulos moved this task from Ready To Go to In Progress on the Machine-Learning-Team board.

Change #1025805 merged by Ilias Sarantopoulos:

[machinelearning/liftwing/inference-services@main] revertrisk: update locust results

https://gerrit.wikimedia.org/r/1025805