How to execute multiple k6 scripts via single command in loadimpact/k6? - performance-testing

Question

As indicated in the official loadimpact/k6 documentation, we are able to execute a single k6 script as follows:
k6 run ../tests/http_get.js
How would I go about executing multiple script files in a single run? Specifically all scripts that reside in a given local directory. Something like:
k6 run ../tests/
Is this supported out of the box by k6?

Answer 1

Depending on your setup there are a couple different ways you can solve this. A pretty straight forward way is to fork the k6 run command inside bash.
#!/bin/sh
k6 run test1_spec.js &
k6 run test2_spec.js &
k6 run test3_spec.js
You could easily write some more complicated bash scripting to read in everything from the /tests/ directory and run them like that. I chose to do it like this though because I had some custom input params to give to each specific test.
Another way would be to write a docker compose script to do pretty much the same thing. This would start up a docker container for each test and run it inside there. The k6 docker image is nothing more than a tiny linux image with the k6 binary added to it.
version: '3'
services:
k6_test:
image: loadimpact/k6
container_name: test_k6
volumes:
- ./:/specs
command: run /tests/test_spec.js
ports:
- "6565:6565"
k6_test2:
image: loadimpact/k6
container_name: test2_k6
volumes:
- ./:/specs
command: run /tests/test2_spec.js
ports:
- "6566:6566"
Both of these methods should allow you to run multiple tests at the same time in a CI environment as well as on your local machine.

Answer 2

At the moment, k6 only accepts one script file, and it runs the exported default function.
import {sleep} from "k6";
import http from "k6/http";
export default function() {
http.get("http://test.loadimpact.com/");
sleep(2);
}
Perhaps, you can accomplish your goal by using modules.
Splitting your logic into modules helps to organize your code and allows reusing your common use cases in different tests.
Check out the k6 Modules documentation
import {sleep} from "k6";
import mainPageUserFlow from "../cases/main-page";
import billingUserFlow from "../cases/billing";
export default function() {
mainPageUserFlow();
billingUserFlow();
sleep(2);
}
Additionally, you could also change the execution of the different Virtual Users on your script like https://community.k6.io/t/how-to-distribute-vus-across-different-scenarios-with-k6/49

Answer 3

Using & will run the tests in parallel, if you want to run sequentially and retrieve the combined result, I suggest:
exit_c=0
(
k6 run script_1.js || exit_c=$?
k6 run script_2.js || exit_c=$?
...
k6 run script_n.js || exit_c=$?
exit $exit_c
)

Answer 4

You can set different scenarios and point to different files containing the script to run each one, as per the docs here https://k6.io/docs/using-k6/k6-options/reference/#scenarios
Here is an example:
import { default as firstScenario } from './firstScenario';
import { default as secondScenario } from './secondScenario';
export const options: Options = {
thresholds: {
http_req_duration: [`p(99)<${httpReqDuration}`],
checks: ['rate>0.80'],
},
scenarios: {
scriptAuthenticatedScenario: {
exec: 'myFirstScenario',
executor: 'constant-vus',
vus,
duration,
},
scriptUnauthenticatedScenario: {
exec: 'mySecondScenario',
executor: 'constant-vus',
vus,
duration,
},
},
};
export function myFirstScenario() {
firstScenario();
}
export function mySecondScenario() {
secondScenario();
}
export function handleSummary(data) {
return {
'results/scenarios.html': htmlReport(data),
stdout: textSummary(data, { indent: ' ', enableColors: true }),
};
}