On localize_server 3.0.x l10n_server_parser queue run is very slow

Made some extra test on gitpod, initial parse from recreate.sh import ins approx. 5mins,
but after that a regular
drush queue:run l10n_server_parser
that is not limited to a very shortlist of release takes forever.

After 8 min running here is where we stand:

gitpod@localizedrupalorg-web:/var/www/html$ drush queue:run l10n_server_parser
 [notice] Retrieving drupal-9.5.0-beta2.tar.gz for parsing.
 [warning] File /tmp/drupal-9.5.0-beta2.tar.gz already exists, deleting.
 [notice] drupal-9.5.0-beta2.tar.gz (7828 files, 10553 strings) scanned.
 [notice] Retrieving drupal-10.0.0-beta2.tar.gz for parsing.
 [notice] drupal-10.0.0-beta2.tar.gz (6783 files, 10075 strings) scanned.
 [notice] Retrieving drupal-9.3.22.tar.gz for parsing.
 [notice] drupal-9.3.22.tar.gz (7774 files, 10462 strings) scanned.
 [notice] Retrieving drupal-9.4.6.tar.gz for parsing.

First time setup will take some time.
recreate.sh currently downloads and parses 14 Drupal core releases.

Just to get a feeling: there are currently ~170116 releases of Drupal core and contrib.
If parsing each will take 2 min per release, it will take us ~230 days to recreate everything.
But recreate.sh only parses Drupal core, which is much bigger than most of contrib.
Nonetheless we are probably better off migrating things.

Quickly looking at the releases since 2005, it's ~25 releases per day.
With a peak of 146 published releases on 2020-06-03.

We will have to support parsing ~150 releases per day on production :)

Thanks @sanduhrs. I understand we'd rather migrate l10n data. Now, regarding onboarding instructions, should we make this step optional, or rephrase it, for newcomers who might only work for a few hours during a sprint for instance?

fmb → credited teebeecoder → .

Make this step optional and improve documentation to explain how to import projects and strings.