Allow bucket specific metrics to find misbehaving applications

__Proposal:__
Add bucket id (like org_id) as label to prometheus metrics (/metrics endpoint)

__Current behavior:__
We have several applications running against one influxdb. Currently prometheus endpoint does not separate metrics based on bucket (or token). If one application misbehaves (e.g. generates high cardinality) it is hard to find out which one through prometheus monitoring endpoint.

__Desired behavior:__
I want to differentiate how much traffic / requests / bytes / errors happen during requests to a specific bucket.

__Alternative behavior:__
We could solve this by creating orgs for every application but this seems like a bad workaround.

__Use case:__
It improves the monitoring of bigger setups of influxdb.

I could prepare a PR for it if desired.

Start at the /metrics endpoint and inspect how Prometheus metrics are currently separated across applications. The work is done when traffic, requests, bytes, and errors can be differentiated by bucket id, with the behavior verified through the project's relevant metrics tests.

Written by the indexing model from the issue text.