[Heartbeat] Shared chrome browser for synthetics

[andrewvc]

We could significantly reduce the memory requirements for on-premise nodes by sharing a chromium instance between concurrent tests. As-is memory usage is quite spikey since we use a full chromium process per run. We could easily share that instance using npx playwright-start or add a npx @elastic/synthetics server command to start a websocket server.

The following benchmarks show that running multiple tests concurrently using a shared chromium is significantly more economical than running a single test with its own isolated chromium instance.

TL;DR: It's close to a 40%~ reduction in memory to use a single chrome instance, saving hundreds of megs per test runner

See bottom of issue for sample scripts.

Results

Scenario 1: Run 5 clients against WS server, then 1 client

Five runners, total memory while connected to WS

Total peak runner mem=1.9GiB
Total peak server mem=1GiB
Mem per run=580MiB

Total peak runner mem=390MiB
Total peak server mem=520MiB
Mem per run=910MiB

Single runner, total memory while connect to WS

Chromium Websocket server while running 5 clients, then one client

t

Single client, no websocket server

Total memory=790 MiB

Reproducing

Save the following runtest.sh script that can run 5x simultaneous runners

#!/bin/bash
set +x
echo Will run $1 times
ws=$(cat ~/.ws-endpoint | jq -r .wsEndpoint)
for i in $(seq 1 5);
do
 npx @elastic/synthetics --ws-endpoint=$ws journeys &
done
wait

Start the chromium server

npx playwright-start --headless

Monitor its memory

psrecord --include-children --plot plotserver.png <PIDOFSERVER>

Run the runners:

psrecord --include-children --plot=plotclients1.png ./runtest.sh

[elasticmachine]

Pinging @elastic/uptime (Team:Uptime)

[emilioalvap]

@andrewc ++, great poc! I'd love to see this implemented, it would reduce hardware requirements by a mile for concurrent monitors

[botelastic]

Hi!
We just realized that we haven't looked into this issue in a while. We're sorry!

We're labeling this issue as Stale to make it hit our filters and make sure we get back to it as soon as possible. In the meantime, it'd be extremely helpful if you could take a look at it as well and confirm its relevance. A simple comment with a nice emoji will be enough :+1.
Thank you for your contribution!

[andrewvc]

Unlabeling as stale 👍

[botelastic]

Hi!
We just realized that we haven't looked into this issue in a while. We're sorry!

We're labeling this issue as Stale to make it hit our filters and make sure we get back to it as soon as possible. In the meantime, it'd be extremely helpful if you could take a look at it as well and confirm its relevance. A simple comment with a nice emoji will be enough :+1.
Thank you for your contribution!