Enhancements to logging when backends are down

[rhysmeister]

I've been testing a setup of InfluxDB/InfluxDB-Relay. It tolerates reboots of instances very well with us experiencing 0 lost records or inconsistencies in our data so far.

A nice enhancement would be to the logging...

1. When a backend is detected down & comes back online.
2. Record count and data size of buffer-size-mb consumed. (Useful for estimating how much node downtime we can tolerate).

[lucas-santoni]

Vente-Privee's fork implements logging feature : https://github.com/influxdata/influxdb-relay . Did you have the chance to try it ? We'll appreciate your feedback ! ;)