feat(decide): Add proper analytics for decide requests

[neilkakkar]

Problem

We want to capture analytics for decide requests so we can charge for these.

We can't individually capture every request, because not only will that be very expensive, but also slow & almost unfeasible given the query volume. So, we aggregate per team:

1. Redis has a dict for every team, a bucket that increments counters for number of requests coming in. Every 2 minutes, we create a new bucket.
2. A background celery task reads from these buckets, and sums up count of all events in all buckets, upto the last bucket; and sends this as an event to posthog.

We do things like this so everything works in a distributed setting: There's a lock on the background celery task, so events aren't double counted (tests show this). While at the same time, this lock does not prevent request counts incrementing in the newest bucket.

We skip the newest bucket for the background task because the numbers in here are still increasing, and it wouldn't make sense to delete this bucket while there are still requests coming in for this bucket.

---

Now, once I know this works okay, I can add to the organisation usage report a query that sums all these counts per team.

One gotcha here is that since we're gathering data from both US and EU, we can't rely on the team_id, as it will duplicate between EU & US. Thus, we are using the team uuid.

---

I've also added an env var for now, so we can de-risk this going to prod. I'll switch it on only in dev first, see everything works fine, then go to prod.

Changes

👉 Stay up-to-date with PostHog coding conventions for a smoother review.

How did you test this code?

[posthog-bot]

Hey @neilkakkar! 👋
This pull request seems to contain no description. Please add useful context, rationale, and/or any other information that will help make sense of this change now and in the distant Mars-based future.