Metrics for monitoring endpoints

[51-code]

Fixes #8

Added metrics for:

• Latency per partition (event's enqueued time - read time)
• Depth per partition (last enqueued event's offset - accepted event's offset)
• Data depth estimation (average message size / amount of messages passed)

Dropwizard setup was inspired by CFE-35 Router.java

Needs to be tested, but this is the initial code.

UPDATE: Tests done!

[kortemik]

@StrongestNumber9 please check this out

[StrongestNumber9]

Added some notes

[51-code]

Discussed with @kortemik that an EventContextFactory can be implemented for unit testing. When producing new events with the factory, we can then call EventContextConsumer's accept() function and test the metrics by getting the Gauge's and asserting the correct value of getValue(). Later the factory can be used for testing the class's other functionality as well.

I'll start working on developing this and I will add it to the PR when ready.

[51-code]

Added testing for the metrics and fixed the issues that came up in the first round of review.

Made a fake object for the Output to get rid of dependencies of the EventContextConsumer object, and the class was refactored with a few new constructors to allow this fake to be injected. Fakes were also done for multiple objects in Azure library for creating "data" for the object, the CheckPointlessEventContextFactory is responsible for data creation.

[StrongestNumber9]

Some minor tweaks

[51-code]

Ready with all the required fixes. @kortemik @StrongestNumber9

[StrongestNumber9]

lgtm at this point