Support multiple workers with new indexing

[ejizba]

Related to #638

This PR does a few things:

1. Always include worker id when logging that we received a request. This makes debugging the "multiple worker" scenario so much easier
2. Use function name as function id instead of a random guid. This ensures each worker returns the same thing
3. Allow the library package to specify both the name and id. This just gives us more flexibility to fix things on the library-side instead of worker-side if necessary down the line
4. Adjust the worker indexing booleans on WorkerChannel for two reasons:
   a. To account for the fact that only the 1st worker gets the FunctionsMetadataRequest
   b. To more explicitly separate the worker-indexing and legacy scenarios and avoid any mingling