I'm starting a project in this realm right now, though only three sensors to begin with. Generally I'm leaning towards "everything in Postgres", but I think I'm going to store the raw sensor data in the filesystem.
I did a set of benchmarks recently for multi-dimensional scientific sensor data. You most definitely don't want row per measurement in PostgreSQL, but you can get surprisingly good results where you store a block of results per row in an array. For even better results TimescaleDB and ClickHouse achieved approximately 2-6 bytes per float32 timestamped measurement, depending on the dataset and shape.
I intend to use TimescaleDB (+PostGIS) for all other data and last year researched this approach for the raw sensor data. I think they have improved compression since then.
For my use, the raw sensor data should only be retrieved 1) to perform a windowed analysis (the results of which will be stored in PostgreSQL) or 2) to display for the user. I'm planning to archive the raw sensor data in files for archival, so I think it's easier to just jam the data in 15-minute netCDF files and call it a day. Will definitely keep an open mind.
This is what I'm doing right now for my home sensor network.
3 ESP8266's with temperature, humidity, and light sensors sending a reading every second to a python app that writes a row to postgres on a raspberry pi 3.
So far the hardest bit has been getting all the services to restart on pi restart. Postgres works just fine.
I'm going to be using an RPI as well and have been messing with systemd. I know people have opinions, but I don't and it was straightforward to configure a Python script to run as a service.
