Description
I have seen a couple different ways that statsd-to-X bridges enabled support for tagging. This project does it by looking for tags embedded as part of the metric name. It is ugly, but probably safest since it makes no assumption about how the statsD process looks at data after the type constant.
A cleaner way is how DataDog does it (see format). I have used this
fork of a popular statsd client library to push data in this way, but it may assume that the DD agent is the statsD service (rather than a generic statsD service that would barf on the data after the type data).
Just tossing ideas out. We are looking at using influxDB and need a statsD-with-tags solution. I can't offer to help (yet), but maybe someone else can get a jump on things using either of these approaches as a basis.