Monitoring PSU usage (Corsair Link)

I have a Corsair RM750i powering my TrueNAS install, and since the latest update changed apps to use docker, I thought I’d try to get measurements from the power supply working.

To that end, I have simply installed the latest version of liquidctl through this docker container avpnusr/liquidctl.

The only “extra” thing I had to do was to check the privileged flag when installing the container, since it needs access to USB devices.

I then attached to the container and ran liquidctl status:

root@0165498a6fd9:/# liquidctl status
WARNING: some attributes cannot be read from corsair-psu kernel driver
Corsair RM750i
├── VRM temperature            53.2  °C
├── Case temperature           47.2  °C
├── Fan speed                     0  rpm
├── Input voltage            115.00  V
├── +12V output voltage       12.03  V
├── +12V output current       12.00  A
├── +12V output power        140.00  W
├── +5V output voltage         5.00  V
├── +5V output current         6.81  A
├── +5V output power          34.00  W
├── +3.3V output voltage       3.33  V
├── +3.3V output current       2.56  A
├── +3.3V output power         8.50  W
├── Total power output       174.00  W
├── Estimated input power    194.00  W
└── Estimated efficiency         90  %

Would you look at that! Everything is here and working beautifully!

  • Don’t mind the Fan speed: 0 rpm, that is expected, the fan shuts off when there is low demand.
  • 140W on the 12V rail is expected as I have 9 12Tb HDDs and am currently expanding my zpool

Anyway, this is cool and everything, but I’d like to get these stats being logged or displayed somewhere - TrueNAS dashboard would be great, Netdata would also do in a pinch.

How can I hook this information up better? Having a semi-idle docker container just to gain access to liquidctl is also kind of jank, but I’m not sure how to do this better.

I know how to do this in Home Assistant, but TrueNAS is not home assistant (duh) so I’m stumped.

I appreciate any help, and in any case this is not critical, I’m just curious.

1 Like

Someone has built a prometheus exporter for liquidctl:

That’s what I’d start with. Using that, you can access the data via HTTP. It’s in prometheus format, which is widely supported by diferrent metric collectors.

From that, you could go with running prometheus & grafana. Or any other dashboarding stack that supports prometheus endpoints. But running,configuring and maintaining those things isn’t always easy.

Might not be worth the effort for just to monitor a single PSU. But maybe you have multiple devices you want to monitor, maybe add some centralized logging collector in the mix. Then going that route might be worth it.

About integrating it into TrueNAS UI/netdata instance… those things aren’t intended to be extended by the user. That said, there is a netdata plugin for prometheus endpoints. But it’d be an unsupported hack to get that running.

1 Like

Run HomeAssistant in a VM on TrueNAS? There’s a guide for that.