I use Amazon Alexa without Nabu Casa but with the “Alexa Skill”. That means I created an own Amazon Skill based on the official documentation.
In general, it works. But when I don’t use it for a couple of hours (like in the morning), the first and often second command doesn’t work. Alexa says the device is not responding and asks me to reactivate the skill in the app.
Within the Alexa app I see “Server not responding” on the devices. But when I try to toggle the switches, after a few seconds they’re back online and everything works fine again. For a few hours…
I’ve seen a similar issue that’s already resolved. But it seems to be related to Nabu Casa (which I don’t use).
So, any ideas how to resolve it for me or how to debug? I can’t see any specific errors in the logs, just the already known Invalid access token (which shouldn’t be the issue as I understand).
Some more info:
Alexa running on Sonos devices, no Echo devices available.
UniFi APs and Switch.
HA running on Synology as a virtual machine (latest version).
Port 443 routed to 8123 on Synology Reverse Proxy, which forwards it to http 8123 of HA (which adds the SSL certificate).
Hi Viktor, this happend to me. Since last month everyday.
When I try to turn off lights I must tell Alexa 3-9 times the same thing and sometimes Alexa tell me to try reconfigure skill again… But I dont know why in some time later, alexa lets me to turn off/on the lights.
Kind of. It seems to work better with the firewall disabled, but I don’t know what’s blocking. And I don’t want to leave it off all the time
So, I’ll have a look in a few days when I have more time
Just jumped on the HA train (coming from Node Red). Setup (including) Alexa was lightning fast - used the self-hosted Alexa skill version (like above).
And: have the exact same problem. Mostly in the morning (or sometimes during the day) Alexa tells me that my device is not responding and I should try re-installing the skill. Second or third try solves the problem and each call thereafter is successful and fast - until this timeout (or whatever it is) happens again.
Looked at all the logs. It seems that the skill is not even reaching the lambda function (no log entries in Cloudwatch). And no log entries on the HA side as well.
So I suspect the skill itself not working properly - but have no clue how to debug this or where to change a setting that could have a positive impact. De- and re-activated the skill a couple of times - no effect.
Firewall (as suggested above): my HA is behind an NGINX reverse proxy. HA itself is running on a raspi v5 and is pretty fast - network has no lags as far as I can see.
Any ideas? Can’t imagine that I am the only one with this problem.