Hi everyone, I'm new to Hubitat (C7 on v2.3.0.124) and I'm trying to set everything up from scratch. I've plugged in the app for the first time yesterday and have added 7 devices to it (4 dimmers and 3 switches) all of which are GE/Jasco devices. The pairing went smoothly, but what is concerning me is the "Severe hub CPU load" flag that is on the top right notification area.
I only have the devices added. No automation or custom apps, and the device handlers are the built in versions. Would anyone know why the 7 devices would cause the Sever hub CPU load warning?
Welcome to HE and to the online community! That does seem a tad strange for a new, out-of-the-box hub, but there are a couple of things you can check.
If you go to the logs page and then look at device stats and app stats it will tell you what is consuming CPU, if anything. Does anything pop out? (If you actually have no apps installed yet you can focus on devices.)
Are your Jabsco devices zwave or zigbee? If zwave there are some additional things we can take a look at but try the following first.
You can also insall a cool driver by @thebearmay called Hub Information Driver that reports various hub performance stats so you can see it in real time... here's a link if you want to give that a try. (Hint: loading Hubitat Package Manager first will help both with Hub Information Driver and most any other custom app you care to install.)
I would do a shutdown from the Settings tab. Unplug the hub and wait 30 seconds and plug back in. Assuming these Jasco devices are Zwave, that can clear out the Zwave radio which for some reason might have got into an odd state with pairing these devices.
Brad's advice above is also good, the Device stats and App stats may show you (or us if you post pictures) something in particular chewing up processing time.
Brad was also onto something where if a Zwave device doesn't pair correctly, and you pair it multiple times you can end up with ghosts. It might be good if you can post a screenshot of your Zwave Details page (Settings tab, Zwave Details menu) to see if you have ghost or incomplete pairings.
Sorry for the delay in response and thank you both for your advice. I have Z-Wave devices and have installed the package manager as recommended. I will be installing the hub monitor in a bit after I restart the hub as advise by @neonturbo.
I also went to the app and device logs as recommended by @brad5 (images attached) but I'm not seeing the % of busy. I'm just seeing the Time, ms. Is there a setting I have to turn on to see this? It doesn't look like the ms value is that high so I'm not sure if this would cause the high cpu load. Any other recommendation would be greatly appreciated. Thank you again.
On the app and device stats I forgot the default view is what you're seeing. You can expand the field selector and add more. But I agree - nothing stands out. I'd suggest if the soft reset does not resolve the issue we need to engage support.
Thank you Brad for the guidance. Other folks have recommended that I exclude everything and re-add the devices slowly to confirm they aren’t ghost devices. I have only one added in and hasn’t seen that error yet so fingers crossed that this result continues. In terms of adding should I be using S2? The print comes up when I add a device whether I should do a secure connect but I skip it, so I’m wondering if that causes the device to chatter too much.
Thanks again
Perry
For switches and dimmers, S2 is optional and you should not see any ill effects from skipping it. (The pairing process will take a little longer if you pair with S2.)
Thanks for the clarification. I know s2 is recommended/required for locks and such. I just didn’t know if they would be causing a issue with the cpu load if I skipped it. Thanks again for addressing that question. I’ll begin adding devices back in and we’ll report back.
It does add some additional traffic and if you’re re-pariing the devices you might as well skip it but I don’t think that alone would cause what you’re seeing.
Hi everyone, it's been a couple days and I did as recommended by the other poster by removing devices and added them back in one by one. I seem to have the same situation where my five devices are triggering the Severe hub CPU load notification. It doesn't seem to matter the number of devices I have (I removed all but one GE Zwave switch) and it gave the same issue. Does anyone have any other advice on how to resolve this? The latest logs are as follows:
Thank you for checking in @brad5. I provided them with my operating system, email address and Hub UID yesterday morning when they requested but I haven't heard anything back yet. From your experience, how long does support normally take to respond? I'm sure they are busy tackling questions from other new users
I have a little different perspective on this -- I'm quite sure there's been a regression in the Habitat OS. I had a bunch of rules defined and hadn't touched them in about a month. About 2 weeks ago, an update came and my hub has been freezing up 1-2 times a day requiring a power cycle to recover. Just today I've now seen the high CPU warning as well.
There has not been another update since the problem happened, and I kept thinking they must know about it given how severe it is. Apparently not?
That's odd... the last update was 2.3.0.124 and that was in December. In my experiece it's been quite stable. I've found in general where there' a widespread problem that affects a lot of users HE is really quick to address it, often in the same day it's reported. Is it possible you received a beta version?
The last time I reported an issue it was a pretty critical one and they responded within an hour, with a fix available the following day. But that was something that affected pretty much every user. I'd give it a day or two and then maybe PM BobbyD with the ticket info and a link to this thread so he can see what you've already tried.
I'm also gonna tag @thebearmay - he's good on hub performance issues and he actually wrote the hub info driver. Maybe he has some more guidance.
As @brad5 said, the last non beta update was in December. It’s been a couple weeks since the first beta for the next update was released and I didn’t have any breaking changes with that either. It’s been about 2 1/2 years since I’ve had any major problems with my hubs. I’ve found that most of the problems are caused by device behavior outside of the hub’s control. Poorly behaving devices, or router/switch can use up the hub’s cpu time, memory, and radio bandwidth, especially with z-wave since it has far less bandwidth to start with.
Aft'noon. A few questions to start. What HE version are you running and what Hub Model? Just a guess at this point, but if nothing else has changed, the symptoms sound like you may have a device going bad and may possibly have, with the freeze ups requiring power downs, some corruption.
