Ref(CI): Reduce processor/ram used by jest test locally.

[lucas-zimerman]

This changes only impact local tests and not github tests.

If your device has too many CPU cores, our tests will spam too many node processes, making the test to use a lot of memory and CPU resources.
By limiting the amount of workers to 25% I was able to reduce locally the time to finish the tests in 16%, reduce memory usage by 54% on an environment with 32 CPU Cores and 32 GB of ram. Assuming that our GitHub actions are running with 2 core and only 8 GB of RAM, I expect those values to be even better.

Here is my local-measured data given the total amount of workers

Workers : how many workers jest can make
Duration: time to complete the test
Memory Avg: the average memory increase before - during the tests

Workers	Duration Avg (s)	Duration Improvement %	Memory Avg (MB)	Memory Improvement %
100%	10.3067	0%	6167.83	0%
50%	9.176	11.00%	3954.21	35.91%
25%	8.6813	15.79%	2857.76	53.67%
20%	8.984	12.82%	2236.91	63.69%
10%	9.3193	9.62%	1671.26	72.91%

#skip-changelog.

[lucas-zimerman]

Not sure if it's worth adding in the project but here is the measuring script that I used

#!/usr/bin/env node
const { spawn } = require('child_process');
const os = require('os');

function getUsedMemory() {
  return os.totalmem() - os.freemem();
}

const startMem = getUsedMemory();
console.log(`Starting memory usage: ${(startMem / 1024 / 1024).toFixed(2)} MB`);

const startTime = Date.now();
let peakMemIncrease = 0;

const jest = spawn('npx', ['jest', '--maxWorkers=8'], { stdio: 'inherit' });

const memoryInterval = setInterval(() => {
  const currentUsed = getUsedMemory();
  const increase = currentUsed - startMem;
  if (increase > peakMemIncrease) {
    peakMemIncrease = increase;
  }
}, 1000);

jest.on('close', (code) => {
  clearInterval(memoryInterval);
  const endTime = Date.now();
  console.log(`\nTest duration: ${(endTime - startTime) / 1000}s`);
  console.log(`Peak memory increase: ${(peakMemIncrease / 1024 / 1024).toFixed(2)} MB`);
  process.exit(code);
});

All I had to do is place it on the core folder, play with the maxWorkers value and replace the test:sdk script for node run-jest.js

[github-actions]

iOS (new) Performance metrics 🚀

	Plain	With Sentry	Diff
Startup time	1204.06 ms	1215.67 ms	11.61 ms
Size	3.19 MiB	4.38 MiB	1.19 MiB

Baseline results on branch: main

Startup times

Revision	Plain	With Sentry	Diff
534ba8c+dirty	1225.00 ms	1237.43 ms	12.43 ms
3e0a5f9+dirty	1233.65 ms	1239.10 ms	5.45 ms
eb07ba3+dirty	1214.49 ms	1221.59 ms	7.10 ms

App size

Revision	Plain	With Sentry	Diff
534ba8c+dirty	3.19 MiB	4.38 MiB	1.19 MiB
3e0a5f9+dirty	3.19 MiB	4.38 MiB	1.19 MiB
eb07ba3+dirty	3.19 MiB	4.38 MiB	1.19 MiB

[github-actions]

iOS (legacy) Performance metrics 🚀

	Plain	With Sentry	Diff
Startup time	3744.83 ms	3739.34 ms	-5.49 ms
Size	2.63 MiB	3.81 MiB	1.18 MiB

Baseline results on branch: main

Startup times

Revision	Plain	With Sentry	Diff
534ba8c+dirty	1230.22 ms	1231.18 ms	0.96 ms
3e0a5f9+dirty	1226.94 ms	1230.02 ms	3.08 ms
eb07ba3+dirty	1222.46 ms	1220.37 ms	-2.08 ms

App size

Revision	Plain	With Sentry	Diff
534ba8c+dirty	2.63 MiB	3.81 MiB	1.18 MiB
3e0a5f9+dirty	2.63 MiB	3.81 MiB	1.18 MiB
eb07ba3+dirty	2.63 MiB	3.81 MiB	1.18 MiB

[antonis]

Thank you for looking at this and running the metrics 🙇
The changes LGTM 🎸

Running this locally on a Macbook M1 Pro / 32GB, I didn't notice any difference, but I think we should experiment and see how this performs on CI.

Not sure if it's worth adding in the project but here is the measuring script that I used

Thank you for sharing this. I'm not sure if we should add it but we can have this PR as a reference.

It will only be noticeable on the latest Pro Max where the cpu count is higher than 8 cores

[github-actions]

Android (new) Performance metrics 🚀

	Plain	With Sentry	Diff
Startup time	422.00 ms	513.50 ms	91.50 ms
Size	7.15 MiB	8.42 MiB	1.27 MiB

Baseline results on branch: main

Startup times

Revision	Plain	With Sentry	Diff
3e0a5f9+dirty	379.92 ms	450.96 ms	71.04 ms
eb07ba3+dirty	419.49 ms	482.12 ms	62.63 ms
534ba8c+dirty	472.35 ms	537.31 ms	64.96 ms

App size

Revision	Plain	With Sentry	Diff
3e0a5f9+dirty	7.15 MiB	8.42 MiB	1.27 MiB
eb07ba3+dirty	7.15 MiB	8.42 MiB	1.27 MiB
534ba8c+dirty	7.15 MiB	8.42 MiB	1.27 MiB

[github-actions]

Android (legacy) Performance metrics 🚀

	Plain	With Sentry	Diff
Startup time	427.25 ms	416.49 ms	-10.76 ms
Size	17.75 MiB	20.15 MiB	2.41 MiB

Baseline results on branch: main

Startup times

Revision	Plain	With Sentry	Diff
eb07ba3	470.04 ms	473.35 ms	3.31 ms
534ba8c	484.00 ms	499.93 ms	15.93 ms
3e0a5f9	401.72 ms	394.98 ms	-6.74 ms

App size

Revision	Plain	With Sentry	Diff
eb07ba3	17.75 MiB	20.15 MiB	2.41 MiB
534ba8c	17.75 MiB	20.15 MiB	2.41 MiB
3e0a5f9	17.75 MiB	20.15 MiB	2.41 MiB