I have not had to re-authenticate the account for at least 2 months, I am very happy with this integration and I take the opportunity to congratulate and thank the creator/s for their great work
I deleted the pickle file last night and that seemed to work, however I got the double authentication again today (same thing you guys are seeing with multiple integration instances popping up.
How do you install a previous version?
Also - do you need to worry about all the keys and such as you re-install it? I remember this installation had quite a few extra steps.
Ok. Out of nowhere I suddenly had two dead and one living instance. Restarting HA wiped the two ghosts.
Maybe the issues have something to do with the HA release. I’m still on 2021.1.3
Potentially - I am on 2021.3.2. amazon alexa integration is the only one doing this for me.
I was just having similar issues today and couldn’t get it to authenticate with any method. I removed my yaml config for it completely restarted and added only via the UI and has been good now.
EDIT: I spoke too soon… about an hour since authenticating I am getting notifications that I need to authenticate again
I delete everything Intergration, HACS, *.pickle, everything. Reboot HA. Reinstall HACS but the previous version… Installed Intergration and everything work here again
As an FYI, I just upgraded from 3.8.2 to 3.8.3 (and at the same time HA 2021.3.2 to 2021.3.3). So far it is working fine (also I’m in the U.S.).
1 Like
I had the problem again. Suddenly there is a notfication about need to reauthenticate and in integration there are two Alexa Media Players. One dud and one working. This time I clicked the 3dot menu of the good one and clicked reload. And then the dud ghost went away. No reauth needed. I think this integration has corona
The service this guy gives us is unmatched…
Buy him a coffee please
6 Likes
Same here,
And we really should be buying this guy a coffee each. I think so many of us use this now. I know its become so central to so much I do with the house these days 
1 Like
Thanks for the love.
You should update to 3.8.4. The issue was caused by certain voice commands in the history not providing certain data in a json format. Because we generally treat json decode errors as a login failure, when it would process this command it would force a logout. When we tested the changes for 3.8.3, the testers apparently didn’t have the command in their history. I’m not sure what the specific command was (maybe certain skills?), but it could bite you randomly in the future.
This explains why it was so random because it depends on what commands you’ve used with Alexa.
I’m planning to remove the legacy login options in the near future (probably 4.x.x). I left it in for people who had trouble but it’s a pain to maintain so many options. It’ll also probably be more secure as I’ll stop storing the .pickle file and the username/password. This means the auto relogin will not work anymore but given the oauth token isn’t expiring too quickly, that may be worth it.
Wow, first time in my life I had this:
- Recognized 3 alexa integrations
- Recognized an update available
- Updated alexa and restarted HA
- The clou: No reauthentication necessary. Alexa worked directly like a charm.
chapeau!
Tried latest update 3.8.4 … everything broken. Removed everything and reinstalled 3.8.2 and everything is working again…
Mine quit working after upgrading to 3.8.3 or 3.8.4 as well.
Stuck on this error:
There was a problem
Enter a valid email or mobile number
I just noticed a problem with the next alarm functionality, I think… I use next alarm to start various automations when we wake up each morning; to that end I created a template sensor with this template:
{{ as_timestamp(states('sensor.terry_s_2nd_echo_dot_next_alarm')) | timestamp_custom('%H:%M') }}
This has worked well for many months, but the time changed this past weekend and now if I set an alarm on the templated Echo device (for example) at 6:00am, the time shown by the template is 7:00, 1 hour ahead of whatever time I asked Alexa to set.
I’ve checked the Echo Devices, the server time (I run on Debian Supervised), and the docker containers. All report the correct time, so I think it’s either my template, or the integration.
Is this a problem with my template, or a bug that should be filed on Github?
Thanks in advance for the help,
Terry
This looks like the winter offset according to your time zone.
Maybe alexa changed the time zone after an update?
Just checked, time zone is correct on the Alexa device. It’s definitely related to the time change, at least that’s when it started happening, but everything seems to be correct.
Terry
Hmmm. I use this sensor in order to diplay the next timer in my rooms:
next_reminder_wohnzimmer:
value_template: >
{% set next = states('sensor.wohnzimmer_next_reminder') %}
{% if next != 'unavailable' %}
{% set reminder = state_attr('sensor.wohnzimmer_next_reminder','reminder') %}
{% if (as_timestamp(next) - as_timestamp(now())) |int / 3600 |float < 1 %}
{{ (reminder) }} in {{ ((as_timestamp(next) - as_timestamp(now())) |int / 60)| round(0) }} Minuten
{% elif as_timestamp(next) | timestamp_custom('%d.%m.%Y') == now().strftime("%d.%m.%Y")%}
{{ (reminder) }} um {{ as_timestamp(next) | timestamp_custom('%H:%M')}}
{% else %}
{{ (reminder) }} um {{ as_timestamp(next) | timestamp_custom('%H:%M am %d.%m')}}
{% endif %}
{% else %}
unavailable
{% endif %}
The time is correct.
So your code should not be the issue.
Thanks for that, after seeing your code I checked some things. I still think something is wrong with the integration for the alarm setting. Here’s why; I said the following to Alexa while watching the device in HA:
“Alexa, set an alarm for two pm”
“Alexa, set a reminder for test at two pm”
“Alexa, set a timer for 4 hours”
I should have had “In 4 hours” for each item, however, as you can see from the image, the alarm added an hour.
Terry