Use worker attributes for alerting latency SLO's not being met
Right now, we get alerts for queues for which the runtime of a job consistently exceeds the values defined in: https://gitlab.com/gitlab-com/runbooks/-/blob/86f04fb26e7bea636328743cbc027b8301a6d37b/scripts/sidekiq-worker-apdex-scores.jsonnet#L48
Now that we've added a worker latency sensitivity (or urgency) worker attribute, we should use that to define the threshold on which to alert instead.
The thresholds are being defined in #25 (closed)