[Compute usage quotas] (Part 1) Talk to Prometheus and aggregate metrics

[jnywong]

Definition of Done

• Query and aggregate the following Prometheus metrics according to the quota policy defined
• Determine the usage according to the scope of the quota policy and provide a true/false value for output allow_server_launch
• Test in showcase hub 🎉
• Don't actually raise a server error if server launch is denied but record info in server logs
• Validate with tests
• Deploy to earthscope staging hub

[jnywong]

Points to address

Points to tackle in the follow-up issue

• If a user refreshes hub/spawn page after a server launch deny, then this will spam the server logs with 403 errors. How to handle this?
• If the user is bumped into a different group to increase quota limits, then we need to refresh their auth state at a decent rate for this to propagate to the usage quota system
• When the exception is raised in the pre_spawn_hook, jupyterhub.user emits a log message and juptyerhub.handlers.pages emits a log message. This suggests to me there may be a more elegant way to handle this.