[DOC]: Create cuGraph Service user facing docs

How would you describe the priority of this documentation request

High

Describe the future/missing documentation

• how to install the cugraph-service server in an environment where the server is to be run (ie. appropriate machine/cluster with 1 or more GPUs that is accessible to machines running the cugraph-service client)
• how to start the cugraph-service server in the server environment - this includes details about CLI options, how to expose GPUs, how to start and maintain the dask cluster the server is to use for MG, etc.
• how to install the cugraph-service client in the client environment (ie. the environment where the users' application using cugraph-service will be, which may or may not be the same machine running the server) Note: the client can be installed separately and without the server software installed.

Where have you looked?

https://github.com/rapidsai/cugraph/tree/branch-23.02/python/cugraph-service

Code of Conduct

• I agree to follow cuGraph's Code of Conduct
• I have searched the open documentation issues and have found no duplicates for this bug report

Start by reviewing the cugraph-service directory on the referenced branch at python/cugraph-service. Document separate server and client installation, server startup and CLI options, GPU exposure, and Dask cluster setup and maintenance; the work is done when users can deploy the server and install and use the client from the documented steps.