This week’s goal
Measure the server’s power use properly.
I had been describing it as “low power”, which is fine as a vibe but less useful when the electricity bill arrives with receipts.
What I actually did
I used a plug-in power meter and checked:
- idle watts
- normal Jellyfin playback
- file scanning
- a heavier playback test
- overnight idle use
The server was not using much, which was reassuring. The bigger surprise was how much extra power storage and random peripherals can add when left attached permanently.
This turned the question from “is the server efficient?” into “what actually needs to stay plugged in?”
What broke, or nearly did
My first calculation was wrong because I mixed up watts, kilowatts, hours, and confidence.
Fortunately, spreadsheets exist to quietly expose this sort of nonsense.
What I’m keeping
- Real wattage readings
- Monthly and yearly cost estimates
- A note of idle versus active usage
- Fewer devices plugged in “just in case”
SmallGrid takeaway
Measure before optimising. A small server can be cheap to run, but only if you know what it actually uses and what is quietly tagging along for the ride.