I get updates form my sensors every minute and the system performance data I will capture with telegraph will be sampled every 10 seconds but sent at a lower rate to save network bandwidth.
I know that influxdb does work differently, but i did not know the right terms from the top of my head.
I tried to work through a sample with retention policies and calculating mean values form the influx documentation but failed...
About 3. : I need to decide if it would be wiser to store all temperatures inside one measurement, all humanities in one measurement and so on or if I Store all sensor values from the living room in one, all from the kitchen in one and so on....
I did not decide in what exact order i would poll my sensors yet, but I plan on using a semi centralized homematic system, where i can get all sensor values for all rooms from the hub and distributed WiFi devices running tasmota.
Based on putting everything i can read at the same time into one measurement would not really be a sensible thing to do, since one measurement would contain 3/4 of the flat in one measurement while the rest is distributed over 20 little measurements containing only one sensor.....
The querry builder is the thing graphana uses so you dont have to learn the sql dialect