Add spigo instructions

[codefromthecrypt]

zipkin has self-tracing, which can generate pretty interesting stuff based on what zipkin is doing. Another option is to use spigo, which can simulate larger architectures.

Here's something we might want to add to the readme or a wiki vs expecting people to harvest slides..

http://www.slideshare.net/adriancockcroft/microservices-application-tracing-standards-and-simulators-adrians-at-oscon

The following works after you've run docker-compose up. We could probably contribute to spigo to smooth or lessen some of the steps below.

# get spigo
$ git clone https://github.com/adrianco/spigo.git;cd spigo
# build spigo
$ sh build.sh
# run netflix for 2 seconds
$ ./spigo -a netflix -d 2 -c
# post json_metrics/netflix_flow.json zipkin
$ sh misc/zipkin.sh netflix

One way to reduce steps is to make some script to rebase the timestamps of json_metrics/netflix_flow.json during the curl (in misc/zipkin.sh). Ex if we rebase the stamps to today, then someone can just use the files that exist vs generating new ones (via spigo).

Another option is to make spigo report directly as opposed to writing a file first. ex adrianco/spigo#67