@perry_mitchell Today, I ran a monitored test. I cut the power at 7:38 (visible through the drop in input voltage in white). For 12 minutes the UPS claimed to be at 100% (blue). Within 7 more minutes, it was down to 90%. After another 19 minutes it reached 50%. In total 50% battery charge, gave me 38 minutes.

Load was 80 W on average (about 12% of the UPS rating).

Monitoring ran on Unraid (nut-influxdb-exporter, #VictoriaMetrics, #Grafana) and was shut down by the server at 40% remaining battery level.

The #Cyberpower PR750ELCD contains two 7 Ah 12 V batteries in series. Your APC Easy UPS 1500 has two 17 Ah 12 V batteries in series, which should give you significantly more run time. Of course depending on your load. I haven't found exact specifications, but you should be able to replace the batteries with something like this (40€ per battery) https://hobbyhall.fi/fi/tietokoneet-ja-pelaaminen/tietokoneen-lisalaitteet/ups-laitteet/green-cell-agm09?id=2604838. https://www.youtube.com/watch?v=nZEG_-NgNqg Or go crazy and replace them with 60+ Ah car batteries https://www.youtube.com/watch?v=K_S55zITqxI .

@jake #grafana is quite ok but it's alerting is sooo messy and counterintuitive.
I do prefer #alertmanager and #prometheus alerts. That's stopping me from switching to #victoriametrics (which is more resource friendly both in term of cpu and disk storage).

@algernon well I started pouring metrics into #victoriametrics and trying to get anything comprehensible from #grafana. So far it looks extremely horrible, but not entirely hopeless. 😁
But at least vmalert+alertmanager can do fancy things from the data.

@ebel as standalone program prometheus is fine (though I'm using #victoriametrics which I whole❤️edly recommend) but actually replacing the whole functionality of nagios+munin will be, erm, adventurous. Using munin is pain-free and automagic, not quite true for prometheus....
Nagios is ugly but the dependency logic works for me very well. I still ought to reconfigure alertmanager, someday.