Configuration Advice

1

I have the TIG stack running on a Pi4b 8Gb on a 64Gb SD card.
I have setup Influxdb V2.6.1 to store the data on my NAS with a view to prolonging the life of the SD card.
I have data feeding separate buckets as follows:

Iotawatt
Octopus Energy
Solar Assistant
Hildebrand Glow Smart Meter
All running on separate RPi’s.

My aim is to produce Daily, Monthly and Yearly stats for my energy consumption, import and export along with solar production stats.
My problem is that many of my queries are timing out both in Influxdb data explorer and grafana.
I have looked at using notebooks to downsample my data but I don’t get an option to select output to bucket as per step 7 in the instructions.
Am I on the right track trying to downsample my data?
I don’t seem to have any control of the data coming from my sources.
Or am I just expecting to much from my hardware/network?
Would it improve things if I change the data location back to the Pi4?
Or do I need a beefier server to run this on?
Any expert guidance/advice greatly appreciated.

2

Hello @cliver,
Im not sure about the HW @Jay_Clifford would be a great person to ask about that.
As for the downsampling I would try downsampling outside of a notebook with a task instead:

3

I have never used a raspberry Pi but InfluxDB on my experience should manage the data efficiently.

what you described does not looks “Heavy” to me.

How often are you collecting the data from the sources?

How does your queries looks like?

most of the efficiency I have managed to implement on my own setup begins understanding the data schema and the intended outcome… placing the correct aggregate function or ordering the filters, could make a huge difference in performance.

InfluxDB schema design | InfluxDB OSS 2.0 Documentation (influxdata.com)

Optimize Flux queries | InfluxDB OSS 2.7 Documentation (influxdata.com)