Massive Web Server Errors

[ericwass]

I noticed Indigo stopped working with Alexa today and dug into the Event log. There are 1000's of red errors that read:

Web Server Error: OAth service for alexa is temporarily unavailable due to excessive Amazon retries

I restarted the server, deleted the authorization, re-enabled the Alexa Indigo skill, even upgraded to the latest version of Indigo. Still no luck.

I'm running 2024.1.0
An old MacMini 2014 2.8GHz Dual-Core Intel Core i5 / 16 GB RAM
macOS Monterey 12.7.6

Any help would be appreciated.

Thanks

[FlyingDiver]

The reflector server was having some problems earlier today, but I think it's fixed now.

Are you still having problems?

[jay (support)]

Those Alexa errors are transient and shouldn't be logged (the next release will hopefully fix that).

[ericwass]

It seems to have resolved. Thanks

[sparker]

I'm suddenly seeing a flood of these (3-4 per second) after upgrading to 2024.1.

[matt (support)]

Just ignore them for a while – they will stop eventually.

[sparker]

Thanks for the reply. I was just coming back in to say they stopped after about a thousand of them appeared...

Off to solve the Starling Bridge error next (may just involve waiting for the next firmware release).

HAPPY THANKSGIVING!

[JaceJenkins]

Every time I see this in my event log, the Alexa skill unlinks, and I have to go in and deauth Amazon from my indigo account and relink the skill.

This happened last night Dec-8-2024 at around 9:00PM CST.

This seems to happen a couple times a month.

[jay (support)]

As I mention above, those errors are not actual errors, only incorrect logging. It is not directly the cause of your issues. BTW, clicking the Clear Cache button in the Advanced Web Server settings will temporarily eliminate the error, but they will eventually come back - the next release will stop the logging.

This is speculation on my part because we only have a few reports of this and we haven't been able to reproduce or even identify characteristics that those few might have in common (many others, including myself, have reported never experiencing the automatic delinking). I believe it's an Alexa issue because nothing we do will unlink Alexa - that's all handled on their side. When and how they determine that a skill needs relinking is outside of our part of the process.

[JaceJenkins]

Well at first I thought it was my internet connection dropping etc, but my Ubiqiuty logs arent showing any internet drops. I do have two connects but they are in active / passive mode and not being load balanced.

But man its frustrating as it always seems to happen at the worst times, and we use Alexa integration heavily. It's to a point we are considering moving to Siri to see if that solves the issue.

Its just odd, that these non-errors were flooding my event log and I have correlated this a couple times when the Alexa integration breaks.

I love Indigo and am a long time user.

[jay (support)]

They may be tangentially related. Each of those "errors" is a token that Alexa got in the past. The fact that you have a lot of them means that Alexa has spend a good bit of time failing the use of a token, then getting a new one and trying again. It does happen, and the more you use Alexa the more you might see them. If Alexa is having to get new tokens a lot, it could (a guess on my part) mean that it's failing to get a satisfactory response from your Indigo Server.

But again - we do nothing that unlinks Alexa. It's happening on their side and we don't know why it would be doing that. Just as an example, I've never had Alexa unlink itself. It has worked for over a year without having to do any linking.

[JaceJenkins]

I’ll check my ids logs and see if something is happening in there. But the server is a new m2 mini with only indigo on a really clean connection. Not sure what else to check.