RDB Loader: track when a load succeeded

[istreeter]

The Monitoring interpreter is able to track either a failed load or a successful load. But we only ever call it on a failure. We never call the track method if the load was successful.

I assume this is a mistake? If it is deliberate, then I suggest re-define the track method to only accept a LoaderError instead of an Either.

[chuwy]

The whole Snowplow-tracking behavior is big a left-over from very old days

I think apart from what you said already, we should:

1. Track more infroamtion, for example volume, amount of contexts, time loading and shredding took. This is up to debate though because it will be mostly duplicated with RDB Loader: report metrics to StatsD or Stdout #384
2. We can resurrect load_suceeded schema as 2-0-0 with this info
3. On a less related note, we should define a more clear strategy for monitoring and failure handling - right now these methods called somewhat sporadically