deploy Zipkin to PWS

[joshlong]

From @dsyer:

haven't kept a very careful eye on them

​[5:11] 
But it's basically 1. Add spring-cloud-sleuth-stream plus your favourite binder to the client apps, 2. Run the spring-cloud-sample-zipkin-stream to aggregate, 3. Run the standard openzipkin UI and point it at 2

​[5:12] 
It's a bit like Hystrix

​[5:13] 
1. Client apps export via message broker, 2. Aggregator gets the messages, 3. UI for displaying results

and

See "Running samples with Zipkin Stream" in the README

​[5:49] 
It describes how to do it with a local docker-compose.yml but it's the same steps for PWS

​[5:51] 
Just had a look and the only thing I know has changed is that mysql isn't the default store and you have to opt in to make it persist the span data(edited)

​[5:52] 
(So the docker-compose still works but you don't need the MySQL container)

the alternative is to stand it up on AWS and then reference it from there (not desirable, though /build/aws-launch*sh should help there if we go that route.