Alexa Smart Home Skill V3 is now live!

Thank you!

Hi All

I decided to update the Alexa skill code. The one I was using previously was very old. I had no issues following the instructions here and created the ASK-CLI details and deployed the skill. Its visible in the Alexa companion app and I was able to link it via the CLI. I couldnt see any errors, everything looks OK

I forgot all my devices using the Alexa website and ran a discovery.

No new devices found using the alexa skill, only devices from other enabled skills

Any suggestions? CloudWatch shows no errors

I should note its using Login via Amazon to my private openhab cloud server. None of this has changed and worked without issue on the old skill.

Have you gone over the troubleshooting guide? This is the typical syndrome of a bad item configuration causing the discovery process to silently fail on the Alexa side. Staggering is the key in resolving this one.

If we want to increase log verbosity, add environmental variable LOG_LEVEL = info to your Lambda function configuration.

Hi Jeremy I did yes, but given the items were being discovered by previously working skills I didnt think there could be an issue with syntax, is that not correct?

Thanks for that log info, ive added that as an environmental value and I see this now, authorisation failure

No worries. I assume you know how to resolve this one on your end now?

I dont but I’ll try and find some info online. The credentials in the config.js are correct, what other credentials are there?

Is it located in lambda/smarthome/config.js? Step 2.i of update instructions. The file moved when switching to ASK-CLI deployment.

Yes, the file resides there with the right details, no issues opening the baseURL and authenticating with the details in the file to show all the REST items.

You can increase log level to debug and see the request details made by the skill.

Hi Jeremy

I recreated the Login with Amazon profile and that seems to have done the trick. Thanks for your fast response and help!

hi jeremy

it seems there is some kind of time out issue, never happened on the old skill

I use a private cloud connector, openhab cloud set to notifications & remote, discovery works fine and my nginx is proxy_buffering off

Any thoughts?


Hard to say with the amount of information you provided. There is a lot of different reasons why this would happen especially if you have the cloud connector in front. You would need to do additional troubleshooting on your end. Use a specific request that failed/timed out and follow it end to end checking logs to see if you can determine the reason.

One thing to note is that a more recent version of the skill is using keep-alive sockets. So you need to make sure your NGINX keepalive_timeout setting is more than 45 seconds.

1 Like

Thanks, im using the latest skill so I’ll check that keep-alive sockets information! thanks Jeremy

FYI, NGINX timeout for keepalive is 75 seconds.

I am aware of the default value. I was just pointing out to the recent change that may affect connections when using a non-standard reverse proxy setup.

Hey @noppres123 - did you manage to solve the “x is not responding, please check…” issue? On my setup, Alexa performs the action without complaint 99% of the time but occasionally throws this response.

Unfortunately not. I haven’t found the time to investigate further. Still on the backlog… :sunglasses:

Ah fair enough

Still getting timeouts with the required NGINX settings and the new skill. Ill try and dig into it deeper.

Have you tried to increase your Lambda function log level, as per my instructions, and look if any possible log lines that could explain your issue? Usually, a device not responding error indicates that the updated state wasn’t available when the command was issued. This can happen either if the item you send the command doesn’t report a state or it hasn’t been initialized yet. On rare occasion, it could happen on low powered server (e.g. RasPi) where a command received may take longer to update the associated item state. If this is the case, you can try to disable any of the bindings that you aren’t using to free up resources for your OH server to function optimally.

You mentioned about the Lambda function errors you are getting but you didn’t mention the responses you are getting on the relevant Alexa requests. Depending on the Alexa error message you get back, it could be a network/socket error or just an item state error on the OH side. Make sure to go over the troubleshooting guide as well.

1 Like