Analysis and visualisation of GraphQL endpoint usage data

It would be great if we could analyse the GraphQL queries coming into the endpoint to figure out information from that data (Kibana/Prometheus?).

• the most popular fields
• the least popular fields, including things we're looking to phase out (see #32292 (closed))
• complexity scores and benchmarking - including what might be good candidates for cacheing if they are being requested a lot and/or are expensive.

etc.

/cc @digitalmoksha @reprazent @nick.thomas @.luke @jprovaznik @johnhope