🏗️ Arbitrarily cap unarchiving to 2 workers

[mguidon]

What do these changes do?

Currently, extracting zip files is done with a ProcessPool using all available cpu cores. This can be pretty heavy and, for instance on AWS, can make a cluster node completely unresponsive. I suggest to cap at 2. Once this is in, all dynamic services that consume that need to be updated.

[codecov]

Codecov Report

Merging #2384 (11ca7cb) into master (a18f686) will decrease coverage by 0.0%.
The diff coverage is 100.0%.

@@           Coverage Diff            @@
##           master   #2384     +/-   ##
========================================
- Coverage    74.8%   74.7%   -0.1%     
========================================
  Files         516     516             
  Lines       20033   20034      +1     
  Branches     1971    1971             
========================================
  Hits        14985   14985             
  Misses       4531    4531             
- Partials      517     518      +1     

Flag	Coverage Δ
integrationtests	67.3% <ø> (-0.1%)	⬇️
unittests	68.2% <100.0%> (+<0.1%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
.../service-library/src/servicelib/archiving_utils.py	89.6% <100.0%> (+0.1%)	⬆️
...ce_webserver/resource_manager/garbage_collector.py	72.0% <0.0%> (-0.9%)	⬇️
.../director/src/simcore_service_director/producer.py	61.0% <0.0%> (+0.2%)	⬆️

[GitHK]

please move this value in a constant named something like MAX_UNARCHIVING_WORKER_COUNT

[pcrespov]

This is a library, i would add it optional to the unarchive_dir with a default to 2

async def unarchive_dir(archive_to_extract: Path, destination_folder: Path, *, max_workers_processes: int =2) -> Set[Path]:
# ...
with ProcessPoolExecutor(max_workers=max_workers_processes) as pool:
 # ...

E.g. if used inside a celery worker it we might want to take all processors offered by the container.