Allow collecting and uploading diagnostic reports from Sidekiq

Currently, we generate our performance reports from Puma only, api & web nodes.

This was mainly done to reduce the scope and simplify the validation of the feature.

Sidekiq is way harder to reason about than Puma because it processes such a diverse set of workloads, and each shard is configured differently in terms of resource limits

see gitlab-com/gl-infra/k8s-workloads/gitlab-com!2037 (merged)

After we'll validate that the reporting + uploads work for Puma, we could iterate and extend it to Sidekiq.

This will require changes in config/initializers/diagnostic_reports.rb, our k8s-workloads and in tests. We should also be aware of CPU & mem consumption on Sidekiq nodes.