Understood. Thank you both. I have a graph already setup and tracking through webcore graphs. I know we are beating a dead horse on this memory thing. So, I appreciate the responses
I guess my concern is how low memory needs to drop from a reboot before it stabilizes. Since, performing the rebuild of a couple days ago it seems to be better. But tough to tell after this latest firmware late yesterday which obviously causes a reboot.
During the initial reboot memory jumps up to around 525mb and that very quickly drops down through 400 to 300 to where it seems to stabilize a bit more in the high 200's. But then slowly throughout the day of a reboot hits mid to low 200's. I'd be happy if it sits there for several days (a week) before a reboot. Seems to at least be the sweet spot just from where many others say they are.
Yep.... In the end it depends.... Different people (HE hubs) run different Apps and devices, so the usage pattern will vary wildly across even the most seasoned members of the Community.
While my comment to @rlithgow1 as largely tongue-in-cheek, utilising InfluxDB and Grafana can give you longer-term data to understand how your setup performs. It may not be something you need long-term, but even in the short-term, it could be useful to analyse your hub's normal patterns.
I think the days of having to monitor the hubs free memory and having to reboot might be over for the majority of people
My longest run in Beta was 6 days, you see can see how stable it is here. The C8 hub is running my home, its not totally loaded down, but definitely not idle. Each spike is a reboot and usually an update.
What with the new official release and the beta program, I seem to be upgrading every couple of days, at the minimum. Hard to get any long term results.
Just wanted to come back to this thread and say that a backup/restore cycle solved my problem. Haven’t had memory problems since. Hope this helps other folks!
Hi all,
I recently installed the Hub Information Driver v3 to track my 3 x C-7 / 2.3.8.122 performance as things are shaky recently with some automations working only partially from time to time or not triggering at all etc.
Looking at the data of all 3 x HEs - it seems like all are having memory leaks - free memory is dropping overtime. I did try to rebuild the DB of one of them but with no change - there is still a leak.
Any ideas how to T-shoot it would be high appreciated.
The yellow line is the hub I rebuild its DB but after the reboot there is still constant drop.
FYI; I rebooted my hubs about an hour after updating to the latest firmware. With this update, the free memory is the most stable it has been in a long time for me. There is still some drop over time, but it seems to be more gradual.
yes, in the last few weeks some automations are no longer fully working - which they have being working flawlessly for weeks/months before - now I need to manually restart each of the hubs almost weekly to recover it and have the automations fully working for few days until next time.
I did not manage to pinpoint it yet to the specific root cause as nothing changed recently in any of the hubs, so I started looking at the CPU/MEM/DB to better understand what is causing this behavior...
Unless the hub is below around 150Mb free I highly doubt that is the primary cause of the issue. Memory decrease seems to be a symptom of hub performance degradation, not a cause of it.
I have very good luck with : 2.3.8.138
Have gone 7days now .. most i have ever got without reboot. LOL
I have lots of issues with weird stuff not working. and Mem Leaks.
I did the update to : 2.3.8.138
Then did a full reset and then restore. < seems to have worked ok of me anyway
