Add apdex weight scores to allow us to aggregate apdex scores across multiple Prometheus servers in Thanos
This will help with the migration to sidekiq k8s.
Part of gitlab-com/gl-infra/scalability#351
Additionally, by providing a weight, will help us answer this question from @sloyd:
Is this issue contributing to the continuing problems we see with Sidekiq availability performance (in addition to authorized_projects)? If so that might be another reason to escalate this. We're continuing to see sidekiq drag on the SLA board even after the improvements end of last month. https://dashboards.gitlab.net/d/general-slas/general-slas?orgId=1&from=1588291200000&to=1588809599000
Edited by Andrew Newdigate