Hey Bob,
I attempted to upgrade to ALPHA again today; but I’m back at BETA for now. Influx porting turned out more complicated than I expected to upgrade (also ran into backfill issues, not 100% sure what those are yet. If I figured that out I’ll make a new thread but it may have been related to all my influx tinkering)
I have another couple of comments regarding Influx usage. The regex field based selection you advocate seems to mostly work. The logging of each “unit” is also great.
After some experimentation, I ended up with:
Right_Mains_W
Right_Mains_A
Right_Mains_VA
Left_Mains_W
Left_Mains_A
Left_Mains_VA
...
This pattern, is however, pretty painful to follow for entering in via the UI (I ended up writing a script to generate the JSON) and its even worse for graphing within Grafana.
While I can write a query such as select mean(/Left_Mains/) from measurement where device='iotawatt' that just seems… strangely formed.
Why is “measurement” there? Is this because of the need to support multiple IotaWatt devices (I would expect the “static” tagging to enable that, not the static measurement).
select mean(watt) from Right_Main where device='iotawatt' seems much more “clear” to me. Perhaps the unit can inform what the field will be? And there would be no need for me to duplicate the “output” field for every measurement I was interested in. Unfortunately this construction does make it /impossible/ to actually do math across measurements; Influx’s continuous queries would be needed here.
Inverting this…
select mean(Right_Main) from watt where device='iotawatt'
Rehashing previous discussion in a new form, I’ll advocate a more influx-idiomatic format:
select mean(watt) from iotawatt where device='iotawatt' and output='Right_Main'
or
select mean(watt) from iotawatt_device1 where output='Right_Main'
Each of these allow arbitrary combinations and segmentation based on requirements.