feat: Add EndpointTraceItemStats RPC

[shruthilayaj]

requires getsentry/sentry-protos#101, will bump package once it's merged

For comparison workflows, we need the count distribution data of all attribute/value pairs (both string and numeric, however we'll focus on string first). We're introducing a trace item stats endpoint to facilitate that. The aim is to build a somewhat generic endpoint that can give a high level overview of the shape of your data.

This implementation is fairly simple, restricted to string attributes to begin with. It returns the frequency distribution of attributes in the requested filter. The data returned is ordered by keys that have the highest frequency attribute values.

This first pass aims to also unblock work the ML team is doing to finalize the algorithm to rank these attributes.

[phacops]

Clean up comments that are not useful anymore, the duplicated constants in the endpoint and the resolvers.