add internal event tracing capabilities

[benfleis]

For diagnostics, analysis, and even when ringpop is scaffolding for other tools, it would be beneficial to have a generic mechanism for forwarding ringpop internal/protocol events.

[benfleis]

Address lots of good feedback from @jwolski. Next up,

• tests
• timers (default expire, disallow infinite?)
• tchannel tracer self removal on some (network level?) errors?
• a few useful events

@jwolski, @CorgiMan: what do you think about events. These are a few off the top of my head:

• membership update
• mark (suspect, faulty, okay)
• join (+ recv'd)
• ping (+ recv'd)
• ping req (+ recv'd)
• admin stats snapshot

If we envisioned doing stats via event, could have an admin point to configure how often they're emitted, and then top/etc. could simply subscribe.