Reading the plan for 2.0, I will say this: there is no reason for anyone using influxdb 1.8 (including myself) with grafana to upgrade.
Reasons not to upgrade:
#1 - increased installation footprint - influxdb now comes with capacitor & telegraf. I would strongly suggest using an approach where all there applications are released in lock-step as individual RPMs or DEBs. It is not uncommon for a single application suite on Linux to have many component RPMs, e.g. openssh-client/openssh-server.
#2 - pushing remote protocol agents into telegraf. This increases the footprint of the database from just needing influxdb to also needing telegraf. To me this seems like a regression. It increases the failure points and latency of getting data into influxdb from other agents. This is a massive reason why I use influxdb: installation profile is easier and simpler than other options.
What I see is that you’ve got a well used and regarded time series database (influxdb) and now you’re trying to use that to leverage getting Kapacitor and Telegraf used by people that currently don’t need or desire to use either of them. (I’ve tried Kapacitor, deleted it the same day.)
As long as grafana continues to use InfluxQL, there’s no observable benefit for anyone to use 2.0 over 1.8.2. The query builder in grafana is very useful and being able to edit InfluxQL (being similar to SQL) makes it very user friendly. FluxQL is the complete opposite of being user friendly.