My problems first started when i wanted to update the rooms app. Suddenly the whole hub froze and it stopped responding to anything. After a few restarts i could finaly open de web ui but at a painfully slow speed. So i contacted support and they told me the rooms app and some other things are giving errors and i should update to the new firmware.
Well i updated to the new firmware and left the system alone for a few hours. The system seemed to be fine again, without any problems. Now suddenly at night all my lights went crazy because the hub was responding so slow that the automation got stuck or something. I couldnt open the webui again and couldnt control any lights. I wish support could help me with the errors so we can see what is going wrong (with @bangali). I unplugged the hub for now because i really wanted to sleep.... Anyone else having these problems? Or am i the only one?
Nope -- I have been having similar issues with the hub locking or the hub powering down and I haven't even touched the hub when it happens. I asked @bobbyD to look at my hub but never did get a reply back.
Gentlemen, as promised, I have pulled your logs last night and both of you have significant volume of errors. In particular, @anon61068208 has 60MB worth of errors in the past 3 days. That amounts to over half a million lines of errors spread across multiple files. @mjw-tan's hub is not quite there yet, but is heading fast to that point.
As mentioned to both of you back in September and then again in October, webCoRE coupled with other big apps that are known to cause hubs to become unresponsive, can lock your hubs, which is what @mjw-tan's logs above are showing. As next steps, I strongly suggest removing these apps (the same suggestion I've made to both of you in September and then again in October). Once your hub is stable, you can add one app at the time and work out any errors you see in your live Logs.
Unfortunately, at this point there are too many errors coming from these apps that attempting to solve them all, would require significant time spent by our engineers researching these issues, which I am sure you can both agree that it would be better spent on enhancing the platform.
We would like to see these apps run smoothly on Hubitat Elevation, so let's take it one at the time and work through problems with community developers and let Hubitat engineers work on things that would further the experience of all.
Thank you both for your continued support and understanding.
I would like to chime in and say this has been my exact experience. Initially installed WC + pistons from ST but started having extreme hub slowdowns. Removing it and creating equivalent RM rules/triggers made things work MUCH better.
Would eventually like to see WC more stable on HE but for now am running fine..
Also want to mention sometimes custom drivers can cause issues as well.. had some weirdness when trying to port over ST stuff.
Is there a possibilty to see these errors so we can look at it ourselfs? I based my main automations on these apps and it would be a big hassle to just delete them. I understand that you cant possibly attempt to check all apps but we could ourself so we can try and fix/change the apps.
If I can make a recommendation...rather than spending the time to dig in a troubleshoot webCoRE, why not spend the same time working to convert your automations over to RM? Then you're on an officially supported/integrated app rather than one that seems to need frequent patching to keep up with the platform updates.
Well it's mainly the Rooms app that i'm using. It would take alot of work to get all those functionalities into RM. I just converted everything from Webcore to RM but i still want to use Rooms. If we can see what goes wrong in the logs we can see if we can fix it.
I just used Webcore because i was used to it. Last time i had troubles i deleted an instance of it and left a small instance because it was working fine again.
I installed the Rooms App last night and after adding just one room my hub responsiveness was horrendous. The room I enabled (and I disabled the RM rule) took a couple seconds longer to realize I was in the room. Felt like I was back on ST. But the killer was that it made all of my other RM automation rules just as slow as that one room. I really love the app and would like to leverage it but it is just doesn't provide the same responsiveness as RM, ML or SL. So I deleted it and everything is fast and furious for lighting automation. Bummed as I know @bangali is an incredible programmer and has an incredible application(s). Just this port isn't working for myself.
Now FYI I do run a LOT of other 3rd party developers applications to support my very heterogeneous device environment; so I am not saying don't do Rooms. Just my configuration and then add rooms to it doesn't work. His APIXU weather app is the best out there and I leverage it for all my LUX automation.
thats what i am hoping to get to the bottom of. this is almost the same app that runs on ST though its been extensively modified for performance specific to HE.
off course the app runs in a sandbox. so errors outside of the sandbox are not available to us users. i have reached out to bobbyD to understand the nature of the errors he is seeing in the logs so whatever needs to be done in user apps to fix them can be done.
I have been dealing with the same problem since last week and I have an open ticket with support. I am using the Lifx group of groups app and I have around 6 groups with two of the groups setup as "group of groups". I could be wrong here and may be this is unique to my hub and setup but whenever the Lifx lights were on, I was noticing slow hub performance. Requests to list devices or show a particular device on the hubitat portal were taking anywhere between 45 seconds to 2 minutes. When I turned off all the Lifx bulbs, the hub started responding quickly.
I also noticed the same behavior with Hubi-Poll app. It was setup to refresh about 5 Aeon labs energy switches every 2 minutes. What I noticed is that, if there is a motion sensor event while hubi-poll is refreshing the devices, there is a noticeable delay before the corresponding rule gets executed. We can see this clearly in the logs. This kind of indicates that the hub is probably running out of threads/resources or there is some other contention somewhere that is preventing the hub to respond while some apps are polling/refreshing devices. Again, this could be something unique to my hub and setup. Hoping that this is a software issue that can fixed and not a hardware limitation.
UPDATE: understand from support that some errors from the app may not be shown in the user logs. these errors are still logged to the platform logs which support views.
support has escalated the issue HE dev and they are investigating whats missing from the user logs so app dev has enough information to fix any such error.
Something is going off the rails in one of those apps. It's not a hardware limitation, it's buggy code. We don't support custom apps and drivers. You need to turn to the author of these apps to get some help, as they are the source of your issue.
Well it seems not all logging is shown to the user because i haven't seen any errors in my logging. So we aren't able to pinpoint the problems with the apps.
The only sort of errors not shown in the Logs are things like stack overflow. These errors indicate the software is off the rails, but don't really provide any direct information about how or why. The fact that the software is off the rails is readily apparent without this. When we see millions of lines of error messages -- mostly traces, it tells us all we need to know. At that point, the only recourse is to the app or driver developer to fix their code.
So how can we determine what is wrong with the app or driver if we don't see any errors? Developers can't see what parts of the code are the problem so not sure how we could pin point the problems.
If you are a software engineer you don't need to ask that question. How about putting logging code in your code, to see where it's falling down? You can glean much more information from your own logging code than error logs will ever show you.
I understand! I will create a rule in RM to replace the hubi-poll app. I do want to use Lifx bulbs with Hubitat and Lifx Group of groups driver was the custom driver listed in the Compatible Devices wiki. Any plans for supporting lifx bulbs (local lan integration would be awesome)?