Add python app to distributed tracing sample app

[lmolkova]

We're working on distributed tracing sample app between azure sdk and azure monitor. We want to publish it on Azure-Samples and let users play with it. It's also a reasonable test bed for us to test how instrumentation integrates with typical user apps and OTel ecosystem.

Here's a brief plan (internal only) and draft with .NET and Java

Adding new service

• add HTTP controller, it doesn't have to do anything for now
• add EventHub consumer: pick a typical way to integrate with EventHub
  • e.g. Java supports messaging through Java APIs, maybe there is something similar in Python?
  • or just use vanilla EventHub SDK with handler
  • use checkpointing with Storage
• Add OpenTelemetry
  • setup OpenTelemtey
  • enable applicable set of auto-instrumentation, make sure to enable popular HTTP clients (including one in Azure SDK), web framework that you use
  • enable Azure SDK collection if needed
  • based on configuration (TELEMETRY_DESTINATION env var), configure either
    • Azure Monitor exporter (Python)
    • or Jaeger

Try to be as idiomatic as possible

• write code as users would do it, assume no manual telemetry collection/context propagation is needed
• pick typical (web) frameworks as users would have
• configuration should be done through dockerfiles, docker-compose and .env files only

App Logic

• App should respond to HTTP server calls
• App should consume messages through EventHub and checkpoint them.
• App should also produce and send messages to ServiceBus.

Add app into the draft repo