ma-load-testing

This package is for migraine again load testing

Prerequisites

• Install k6

Usage

Get a list of urls to use from CloudCMS

• Get a representative sample of urls based on total pages per template / total pages

node bin/hydra-load-testing.js load-urls -r -s 200 -o ./urls.js

The -r option enables getting a representative sample

The -s option specifies the size of the total sample, aka how many urls to get back. You may get a few more pages because the script takes the ceiling of the fraction * total nbr pages

The -o option specifies the file path to write the urls to

• Choose how many pages of each type to get

node bin/hydra-load-testing.js load-urls -t article -t author -n 80 -n 10 -o ./urls.js

The -t option specifies the template

The -n option specifies the number of pages to get for that template

If you specify -t params, you will only get pages for those templates that you specify

The -n value corresponds to the -t template at the same index

• For more options see help

node bin/hydra-load-testing.js load-urls --help

Run k6

Ouputting to a json file

k6 run runner.js -e HOST=https://pre.migraineagain.com -e VUS=100 -e SITE="Migraine Again" --out json=output/100.json

Outputting to cloud

• First make sure you have the K6_CLOUD_TOKEN environment variable set. Reach out to Ali if you need it

• Then run the following

k6 run runner.js -e HOST=https://pre.migraineagain.com -e VUS=200 -e SITE="Migraine Again" -e K6_CLOUD_TOKEN=$K6_CLOUD_TOKEN --out cloud

To evaluate results, follow the guidelines provided in the "Expected Range values for K6" section here

The requests/second value described there corresponds to the http_reqs value in the k6 output