Optimize catalog services endpoint

[sanderegg]

What do these changes do?

• reduces the number of calls to the database
• optimizes validation through ProcessPoolExecutor
• catalog to gzip responses when over 500 bytes

gained a few 100ms in the operation, hopefully might help in the 20 users access osparc from the CI case

Related issue/s

How to test

Checklist

[codecov]

Codecov Report

Merging #2313 (bba4593) into master (c74f23a) will decrease coverage by 3.3%.
The diff coverage is 24.0%.

@@           Coverage Diff            @@
##           master   #2313     +/-   ##
========================================
- Coverage    71.7%   68.4%   -3.4%     
========================================
  Files         506     506             
  Lines       19912   19927     +15     
  Branches     1949    1947      -2     
========================================
- Hits        14293   13633    -660     
- Misses       5143    5858    +715     
+ Partials      476     436     -40     

Flag	Coverage Δ
integrationtests	75.3% <ø> (+13.3%)	⬆️
unittests	67.1% <24.0%> (-0.1%)	⬇️

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

Impacted Files	Coverage Δ
...mcore_service_catalog/api/dependencies/database.py	68.7% <0.0%> (-4.6%)	⬇️
...src/simcore_service_catalog/api/routes/services.py	32.5% <15.7%> (+6.0%)	⬆️
...imcore_service_catalog/db/repositories/services.py	30.3% <25.0%> (-0.9%)	⬇️
.../simcore_service_catalog/db/repositories/groups.py	42.8% <28.5%> (-5.5%)	⬇️
...og/src/simcore_service_catalog/core/application.py	86.3% <100.0%> (+0.6%)	⬆️
...es/sidecar/src/simcore_service_sidecar/mpi_lock.py	29.0% <0.0%> (-71.0%)	⬇️
...simcore_service_webserver/computation_subscribe.py	30.0% <0.0%> (-61.3%)	⬇️
..._director_v2/modules/db/repositories/comp_tasks.py	40.5% <0.0%> (-55.5%)	⬇️
...c/simcore_service_webserver/users_to_groups_api.py	46.1% <0.0%> (-53.9%)	⬇️
...ore_service_director_v2/api/routes/computations.py	30.3% <0.0%> (-53.4%)	⬇️
... and 37 more

[pcrespov]

This is definitively an improvement! :-)
Please consider the suggested changes . There are important changes and I also think we should start adding more testing to the catalog

[pcrespov]

shouldn't we have an executor pool per application instead of per request? Did you test the difference?

[sanderegg]

yes. saw nothing, no measureable improvement