Use reactive caching worker for external services
What does this MR do?
It uses a specific worker/queue to process the work that
require external services and use the ReactiveCaching concern. That's because this kind of
work should not be mixed with other time-sensitive work, or work that doesn't request external services.
Service timings can be found at https://log.gprd.gitlab.net/goto/3dbc7c3a365938d1d55c333cf6136185.
Related to gitlab-com/gl-infra/scalability#224 (closed).
Screenshots
Does this MR meet the acceptance criteria?
Conformity
Edited by Oswaldo Ferreira