tests(gatsby): Add metrics output for filter and sort benchmarks

[imjoshin]

Description

We currently have a benchmarks/query-filters-sort tool for running sites under different node size/count and filter/sort conditions. This PR extends this tool to enable CSV logging so we can better visualize the impact of certain changes.

Example usage:

$ BUILD_COUNT=10 OUTPUT=thousand NUM_NODES=1000 FILTER=gt yarn bench
yarn run v1.21.0
$ ./benchmark.sh

Starting 5 builds...
Writing results to ./output/thousand/

12:39:13 - Build #1: initializing
12:39:14 - Build #1: running with pid 3540
12:39:34 - Build #1: finished in 20090ms
12:39:34 - Build #2: initializing
12:39:37 - Build #2: running with pid 8390
12:39:56 - Build #2: finished in 19392ms
12:39:56 - Build #3: initializing
12:39:59 - Build #3: running with pid 13252
12:40:19 - Build #3: finished in 20118ms
12:40:19 - Build #4: initializing
12:40:22 - Build #4: running with pid 18010
12:40:42 - Build #4: finished in 20153ms
12:40:42 - Build #5: initializing
12:40:45 - Build #5: running with pid 22871
12:41:06 - Build #5: finished in 20771ms

Finished!
Done in 112.87s.

The thousand_summary.csv file:

build,time,max_rss,max_vsz,result,code
1,20090,530,33310,SUCCESS,0
2,19392,532,33315,SUCCESS,0
3,20118,533,33311,SUCCESS,0
4,20153,528,33310,SUCCESS,0
5,20771,524,33310,SUCCESS,0

And each thousand_build_X.csv file:

time,rss,vsz
111,48,574
221,63,588
330,70,599
440,89,754
549,89,752
661,91,756
773,112,760
886,128,773
997,135,781
1108,147,791
1221,163,808
1332,176,851
...

Documentation

Documented changes/usage in the README

Related Issues

[sc-45733]

[imjoshin]

Closing this in favor of #34810