Read/write signalK data per vessel

Currently, all data is written to a single bucket in InfluxDB, regardless of the vessel. Since the introduction of runtime data source switching (see #53 (closed)), it is necessary to separate data by source (simulator vs. real SignalK data). This separation is also a prerequisite for future features, such as clustering vessels, consitently generate tracks, and synchronizing data via Kafka.

!171 (merged) provides an initial implementation that tags data written to InfluxDB with the vessel ID. This feature requires further testing: while the simulators vessel ID works as expected, the SignalK vessel ID seems to be missing atm.