Seem to be having an issue with Alexa integration as well. Lost my UPS this weekend, took down my primary gateway/internet. Ended up not noticing it for 20-30 minutes, but once back online was unable to issue any commands via my Echo devices. Reboot of the HE, and all was well. This evening I rebooted my firewall and ran into the same issue. With a reboot of HE making all well again in the world.
This was the only message found in the logs:
No surprise on the message, just didn't recover once all was back online. I'm half tempted to pull the WAN connection, leaving the gateway accessible to see if the issue resurfaces.
For me, I'm not very concerned with the issue. Just need to remember to restart HE after the gateway is restarted, which is normally a rare event. But if the issue occurs with a loss of internet, I could see other folks having some grief if network drops are common.
Do you have a DHCP Reservation for your Hubitat hub’s ip address in your router? I am guessing every time your network gear reboots it may be assigning a different ip address, causing a temporary disconnect between Hubitat’s cloud server and the hub. This could explain what you’re seeing.
One more observation ... My Harmony integration was also broken but didn’t think twice bout it .... post rollback it’s up and running again with no intervention ...
I do have reservations for all of my associated equipment. During the "event", none of the addresses are changing. Just odd that once my gateway disappears things go south. With how my network is laid out, none of the other equipment would have had loss of ethernet connection. All of the supporting network hardware remained powered up. The only link that would have been dropped is the one from my main switch to the firewall itself. Not a VLAN routing issue as I haven't segregated anything....yet. Quite honestly, I'm really not bothered by what's happening, I just would love to understand why. I'm going to do a bit more testing. Leaning towards the loss of the gateway that would start causing some "unreachable" errors for communications. Would have also included DNS failures as a possible root cause, but my DNS server remained active during the second event and should have cached all of the associated hosts.
Edit: Just tried pulling my WAN connection. Received the same error message in the logs, but integration still worked once connection had been restored. Sill leaning towards the loss of gateway being the root cause.
After a platform rollback, I also restore the previous database version that matches the platform version too. Just in case something changed in the database that would make an upgraded database incompatible or less compatible with the previous platform build.
Sudden disappearances from the LAN could be a faulty network cable or an issue on the switch. Is it a managed or unmanaged network switch? If managed, try disabling STP (Spanning Tree Protocol) on the switch if that option is available.
Good suggestion on the cables/switch. I stopped buying cheap network cables many moons ago.
Thankfully, I know why the network had gone down. My first event was due to the battery on my UPS failing, shutting down the unit. The second was when I installed the new UPS.
In retrospect, the battery failure event also caused the HE to lock up. May be time for me to pick up a second HE to goof with for testing. I'm starting to get the stink-eye from the family every time I try to break things. The only commonality I've found so far is the combination of loss of default gateway on the network and (possibly just a symptom more than anything else) Alexa integration.
@chuck.schwer - Can you think of anything in Hubitat Platform v2.1.9 that could lead to Amazon Alexa Skill reliability issues? I don't recall seeing anything in the release notes that stood out as a possible cause. Thoughts?
Note: The issues reported in this thread appear to only impact a small number of users. My system has been fine since 2.1.9, however I don't have more than ~30 devices exposed to Alexa.
Nothing I can think of, there were no changes to the Alexa integration. However if the hub loses connection to our server then that could lead to an issue since any messages from amazon go through our server and get passed to the hub. I would recommend anyone who has rolled back to roll forward again and see if the problem persists, if it does please let me know. I would like to exclude the possibility that there was a hiccup with amazon at the time. We did not see any reports of this with our beta testing either.
@chuck.schwer I tried both versions of 2.1.9 and both of them caused alexa integration to stop. Only works for a while then stops.. have to downgrade to fix the issue.
Blockquote[quote="chuck.schwer, post:54, topic:34952"]
Nothing I can think of, there were no changes to the Alexa integration. However if the hub loses connection to our server then that could lead to an issue since any messages from amazon go through our server and get passed to the hub. I would recommend anyone who has rolled back to roll forward again and see if the problem persists, if it does please let me know. I would like to exclude the possibility that there was a hiccup with amazon at the time. We did not see any reports of this with our beta testing either.
[/quote]
@chuck.schwer Hi there I’ll try again over the weekend... the Alexa integration was partially broken .. adding and removing devices worked perfectly every time and post a reboot or reinstall of the Skill would work for a while then fail .... that’s with about 30 devices and with just 3 ... number of devices did not seem to make a difference .... but at all times I could add and remove devices to and from Alexa !!!!!! So comms to Alexa from HubiTat was ok ... the Alexa to Hubitat comms seems to be the part failing
My harmony Hubitat device failed outright ..... socket errors/timeouts every time I tried to use it ...
@chuck.schwer I've actually been trying to figure out when the last time the Alexa Skill had any kind of update to it. Its been a long time though.
Actually a new convert from other HA platforms.
Anyways I found a bug right away in the Hubitat Alexa integration which I have already bought to the attention of Support. This has to deal with using RGBW devices, in Alexa if you go from a color -> CT (shade of white) you get a "Device is unresponsive" error. You have to pick a different CT then the default one that hasn't been used yet in order for it work. The same is true going from CT -> Color -> CT.
I think it’s something on the system layer .... networking maybe or que management or something along those lines ==== Both the things that broke for me are IP / Network bound
