Given that the house network arrangement worked fine as-was for two months and that Hubitat administrators have not mentioned any service downtime, It is more likely that something happened in the networking code of Hubitat that is not handled and exposed to the system logs. Networking code should, and I expect Hubitat does, pole the socket connection regularly (tight loop - most likely milliseconds tight even), and upon detection of a fault, close and reestablish the socket. What I can't say is what happens if for some reason the dynamic connection settings in Hubitat become corrupt or more likely stale, eg. a bad DNS lookup cached in memory, and the code goes into a loop of trying to connect to the wrong IP. Writing highly resilient code that recovers from or at least effectively reports every possible failure condition is no simple task. The code has to determine it is in a possible failure condition...
What would help is if Hubitat code would log core IP network events to the system log like it does device events, maybe even provide log level configuration to the owner/user. (trace, debug, info...)
These logs would not only help Hubitat, from a code quality perspective but might also help the owner diagnose local/internal and "last mile" Internet network issues. I'm not convinced it wasn't a last mile issue that might have provided bad data (eg DNS) to Hubitat and it simply never recovered short of a full restart.
Providing the owner/user more granular control, eg network restart, service restart would also help triage and identify the real fault cause. Reboot should be considered the diagnostic of last resort. Reboots tend to wipe out all evidence of the real issue. Moreover, when rebooting Hubitat clears the symptoms, the implied diagnosis is that the fault point was within the Habitat device.
Footnote: I am a three decade IT veteran (admin, engineering, architecture, infosec, and more) with over a decade specifically building high performance, highly available network appliances (hardware design, firmware, kernel, OS, and user space architecture and development). My observations and suggestions here come mostly from that over decade period, inclusive of my extensive experience dealing with enterprise customer engagements (eg live product debug in production environments).
Congratulations you'll find many here with equal and greater experience, along with about the same number with lesser to no experience, so we try to strike a balance when crafting our answers. Having said that I'm going to go back and see what has already been discussed before I dive in with my suggestions.
7/30 follow-up. It has been four days since the last failure and the situation appears to have stabilized (returned to normal). From my point of view the immediate incident is "over", but I have no root cause that will help avoid or at least better diagnose a future series-of-failures incident.
I am still open to suggestions and further discussion of triage and resiliency improvement.
If and when the issue happens again I will screenshot the results.
Unless I am in an active broke state (and even then) I am loath to do this as it creates issues on other devices that can and do shift between bands based on active monitoring of network quality. What I will do if it happens again is insert a temporary media bridge that I know has no issues (proven track record) with unified SSIDs.
THis exact thing just happened to me, both hubs , a c7 & 8 AND my HD+ on phone and windows which uses the cloud url failed, all at same time. In my experience this points at the cloud system.
I have a very similar problem... C-8 Pro running very recent version. Works fine on the cloud for weeks, but then becomes inaccessible from the cloud for days and does not correct itself. I have a weekly reboot scheduled to "fix" the problem, but I'd rather it didn't disconnect at all. It's a vacation home, so debugging is hard since I'm not there.
I just wanted to say I've been having this problem for about 5-6 months.
Cloud connection unavailable.
Hub is running latest version but has had this issue over multiple versions.
Its connected via Ethernet cable to a switch that then goes to the router.
I have a very reliable webcore piston that auto reboots my routers should an internet test fail consecutively 3 times over 15 minutes and it is not triggering.
All other systems in my house have no issues.
Manually activating a router reboot does not fix the issue.
I have to connect locally to the hubutat and reboot it to fix the issue.
It seems to occur every 6-10 days. The first indication it's happened is my Google home is no longer able to control my devices.
Router, switch and hubutat all run on UPS'S. I also have a power outage webcore piston that detects brown outs and blackouts and doesn't show anything around the time of the cloud disconnections.
I have a C7 hub.
If I could view hubutat reboot logs I might be able to pinpoint when the problem started occurring. I never needed to reboot my hub until this started happening.
I have performed a hub restore too (unrelated to this problem) and it's still doing it.
I wish people would stop using this phrase and give the version number, as this really doesn’t tell me anything other than you believe you are current. (I’m guessing that in about 15-20% of the cases the OP is wrong in their assumption.)
That having been said the loss of cloud connection is not necessaily a loss of the internet but a notice that the hub is no longer talking to the HE Cloud, and that cloud backups and remote dashboards among other things will fail.
Me too... Almost worthy of a topic / rant of it's own.... No disrespect to those that have used it in the past, more just a pent up frustration I at least should have vented before now...
Hmmm. shame... should be some way forward.... the guys on the front-line have done huge changes in recent months. Not to say they have fixed anything in this space, but it should not be beyond them to fix an issue like this....In My (un-educated) Opinion.... (IMueO)
Copy that, thanks for the response.
Done and done, will report back when it fails again.
I did tap the use router DNS settings but the box is still empty. If that's normal?
