Add [distributed] tracing

Add tracing to the code and create spans in the code. This instrumentation will help us to understand how flintlock is functioning, where the bottlenecks are etc.

We will need to be aware of any existing span ids that may be passed via gRPC to one of our end points and if one already exists use that as the parent. ....basically so distributed tracing works. Ultimately the CAPI provider will be the root of the trace.

We could use OpenTelemetry.

Start by mapping the Go code behind the gRPC endpoints and how requests from the CAPI provider enter flintlock. Determine where existing span IDs can be received and propagated, then define the spans and parent-child relationships needed for distributed tracing; done means tracing covers the relevant flow and preserves the CAPI provider as the root.

Written by the indexing model from the issue text.