Weekly escalating task work

[armenzg]

I'm seeing a bunch of issues here:
https://sentry.my.sentry.io/organizations/sentry/issues/?project=2&query=weekly_escalating&referrer=issue-list&statsPeriod=7d

Getting ahead of it will save us from having trouble down the road.
I believe fixing the slow DB issue would reduce some of the timeouts we're seeing on the cron monitor.

Notes from Josh:

• modify this referrer snuba side to not be a customer rate limit, but a global rate limit based on the referrer (this is what we did for the snuba part of ML backfill)
• determine if we're making the queries too quickly. If that's the case, we can decrease the concurrency of queries, add intentional slowness to the task, and/or try to improve the queries that it makes