Additional lightwalletd metrics and monitoring

[benzcash]

A compiled list of request monitoring that would be useful for wallet developers.

• stats on the range and frequency of blocks served so that we can get a clear idea of our caching needs (how far back is best to cache)
• Average amount of time it takes a user to sync (this may be tricky but if we could try to get some visibility it might help)
• Reorg detection
• Unique users / day / month
• Metrics that help us understand our bandwidth needs
• Cache misses (we could probably infer this based on response times)
• Memory use so that we can get an idea of whether we can increase the cache size. Can we cache the entire chain?
• Frequency of failed sends. Does this spike around reorgs?

Some of this is possible once we get the prometheus framework merge (https://github.com/zcash/lightwalletd/pull/172) and some will require research.

@braddmiller can we create an epic to split these issues up and prioritize based on importance?

[gmale]

maybe it is so obvious that it doesn't need to be enumerated but "downtime" should be monitored so that we can react quickly when the service falls over.