Remove span-normalizer service

[findingrish]

The code in this service is basically simple transformational logic and can be easily moved to either collector or raw-spans-grouper.

cc @kotharironak @jcchavezs @buchi-busireddy

[codefromthecrypt]

wouldn't that improve performance? :P

[jcchavezs]

@rish691 if have time, would you describe exactly what is span-normalizer doing? Moving to collector is also a good idea cc @davexroth

[findingrish]

I am a bit late in replying, sorry for that.

Let me write down what it does:
It consumes the message published by collector, in the following format:
https://github.com/hypertrace/hypertrace-ingester/blob/main/span-normalizer/span-normalizer-api/src/main/proto/jaeger/jaeger_span_internal_model.proto#L88

It then converts this span data into an internal format called RawSpan:
https://github.com/hypertrace/data-model/blob/main/data-model/src/main/avro/RawSpan.avdl#L8

The key difference between the input & output schema is some first class fields https://github.com/hypertrace/data-model/tree/main/data-model/src/main/avro/eventfields
These fields are populated by some if/else logic over the input data attributes.

I feel, running this job as standalone service is not necessary and can be merged with an existing service.