utilities/python_influxdb_downsample#3: Allow multiple outputs to be specified

This one is probably a little more complex than it sounds.

We currently specify the output instance and bucket seperately

        output_influx: home1x
        # Which bucket?
        output_bucket: testing_db

If we're going to add support for multiple outputs, do we want to be able to specify different buckets? That'd be possible if this were a flux query as the bucket is specified when calling to().

The problem is, that quickly gets complex, with a number of failure modes.

I think the answer is probably to say that the same bucket must exist on all configured outputs.

There are a few obstacles to cross here then:

• We need a write client per output
• We need to maintain support for the existing string notation
• Need to decide how to act when one write fails

Not insurmountable, but needs considering.

If we're doing this anyway, do we perhaps want to say fuck it and allow the bucket to be overidden? That'd mean accepting a dict instead of a list

        output_influx: 
             home1x:
                  bucket: testing2
        # Which bucket?
        output_bucket: testing_db

Whilst that feels a little like over-engineering, it's nothing compared to the complexity involved if we decide we later want it and end up having to support string, list and dict.

If we're doing that though, do we in fact want to accept a list of dicts, so that the result can be written out to multiple buckets on the same instance?

        output_influx: 
             - influx: home1x
               bucket: testing2
             - influx: home1x
               bucket: testing3

        # Which bucket?
        output_bucket: testing_db

Again, it's probably overengineering but leave us with the most flexibility if we want to change stuff later