Skip to content

README.md

Latest commit

 

History

History
51 lines (38 loc) · 2.11 KB

README.md

File metadata and controls

51 lines (38 loc) · 2.11 KB

Load Test with Locust

This directory contains a Locust script for load testing.

│   locustfile.py    # gRPC locust client
│   load_test.conf   # load test configs
│   test_grpc_api.py # test script to call gRPC server
│   cat_224x224.jpeg # resized test image

How to run load tests

  1. Install Locust with pip. You have to open up a new terminal session (or sourcing the ~/.bashrc) to enable locust CLI.
pip install locust

  1. Replace placeholderes. In load_test.conf and locustfile.py, there are three placeholders of <<EXTERNAL-CLUSTER-IP>>. You have to replace those with the actual endpoint that your TFServing is deployed on. Also you need to replace <<WHERE-TO-STORE-RESPORT>> placeholders in load_test.conf with where you want to save the final report.

  2. locust the load_test.conf. Every bits of configurations are defined in load_test.conf, so you only need to specify it in --config option.

$ locust --config=load_test.conf

Inside load_test.conf

locustfile = locustfile.py
headless = true
users = 150
spawn-rate = 1
run-time = 5m
host = http://<<EXTERNAL-CLUSTER-IP>>
html = <<WHERE-TO-STORE-REPORT>>.html
csv = <<WHERE-TO-STORE-REPORT>>

More complete descriptions for each configuration can be found in the official doc, but here we provide a brief summary.

  • locustfile: the python script file which implements the client behaviour.
  • headless: whether not to use UI or not. If you set this True, no UI is involved, and the load test runs right away.
  • users: the maximum number of users(requests).
  • spawn-rate: how many users should be added at a time. You can decide the time interval between spawns in the locustfile.py.
  • run-time: how much time the load test should be conducted. Here, m means minutes.
  • host: endpoint
  • html: path where the HTML based report should be stored.
  • csv: path where the CSV based report should be stored. Notice that we don't set file extension .csv here because multiple csv files for each circumstances would be generated.