Setup usage datastore and store usage

[easyCZ]

Depends on #10320

We want to start persisting our usage report such that it can be viewed and queried quickly (from the Public API, and dashboard). In this task, we're defining the datastore and persisting the report into it.

• Create table to store usage data (Add migration for new usage table #11311)
• Persist data on every reconcilation run ([usage] Persist usage reports in database #11343)
  • Populate creditsUsed columns ([usage] Persist credits used for each instance in the usage store #11406)

[andrew-farries]

We decided:

1. Store usage data in a new table in the existing db, called workspace_instance_usage or similar. This will have fields: instanceId, attributionId, startTime, stopTime, creditsUsed and some kind of generationId to show the last report in which the instance appeared.

2. Separately, we would also like to persist the raw usage report (as json), possibly using content-service so that we have an audit trail. The generationId field in the usage table could then be a link to this report.

[geropl]

That sounds good to me. I'm not sure we need 2. (the document store), though, but happy to learn about the motivation once we ticked 1.

Regarding the schema: I did not check all fields, but even given a generationId, a generationTimestamp or so would greatly help with debugging, at the very least.