Those are wild swings in the runtime. There is nothing connected to the UPS that should cause those large swings. (yes, 2x NAS, two macs, a few notebooks etc) but their power usage is not really that valotile.
I set up runtime reporting via Node-red , triggering whenever the reported runtime changes by 10% up/down from a prior reading, and that keeps alerting me every few minutes, with runtimes from 0h45mins to 1h20mins.
Just wondering whether other have seen this and whether it coud be a reporting error, or real varying usage?
Have you run calibration test on the unit at least quarterly? Batteries age over time and the charge/discharge does change the runtime accordingly. Learned that way back in the 90’s for all batteries.
I have the rack mount version of this one and its setup the same, connected through my synology NAS acting as the UPS server for NUT.
I have not had issues with fluctuations in estimated on battery run times. This is what I have setup for my automations currently for it as I am going by battery level not estimated time left:
I have the same or similar UPS. My runtime looks like that. It’s normal. Your NAS power usage will fluctuate as tasks run on your NAS. If you have any backup tasks, data scrubbing tasks, or any application that would cause a high CPU load periodically you will get these fluctuations. My UPS typically sees about a 73W load from my NAS (7% total load on the UPS). When data scrubbing or a backup occurs it will spike to about 90W (9% total load). That 20-25% change in power or about 2% load equates to about a 35 minute drop in runtime. My data scrubbing just occurred a couple days ago and ran from about 1am until 1pm. I then had a TV recording occur around 8pm for a few hours. Here’s the graph showing load and runtime correlation.
This is a brand new battery pack about 1 month old that has had runtime calibration run twice since being installed.