This is for those users in the EU region continuing to have issues... #2483
Comments
|
If it works out that a change in one or more of these works out for EU, I can see if I can make them configurable from the UI... |
|
@danielbrunt57 It seems working here from Italy. I changed all the configuration you suggest |
|
in Italy it doesn't work, the reconfiguration pop up appears continuously
|
|
Updated AMP to 4.2.11 under Supervisor 2024.7.4 and then I updated to 2024.8.2. Domain .es Edited files: File 1: manifest.json File 2: const.py I didn't see error at logs after restarting HA. But after 10 minutes I could see the following warnings. Warning 1: Warning 2: Warning 3: Then I reloaded the integration and I got the following errors at log. Error 1: Error 2: |
|
Hi, same error here, also for the ".es" domain Also getting this error below in the logs: I manually deleted the .pickle file, but it is recreated and the error repeats. |
|
Did not check what was changed, but a reinstall of the previous version saved it immediately. |
It's impossible for me to help you troubleshoot your issue without seeing your related DEBUG logs...
|
|
This is the code in the api _request call: |
|
It's really going to take someone who has the issue and also has somewhat expert debugging skills to get to the crux of the problem... |
|
This is the HA debug log chunk when AMP started malfunctioning, don't know if it helps. |
|
Just for if it helps. I enabled the debug registry from AMP and then I restarted HA. So this is the log since the restart until it fails. I edited serial numbers, email, ... I did the test after updating to Core 2024.8.3, released today. |
|
I tried the suggested workaround changing the const.py, but after the reboot i receive a "bad configuration" notification from HA with this error log:
|
|
Here is a complete debug log |
|
In .es always reconfigure |
|
I just tried again as it works for me. Core 2024.9.0.dev202408260813 Manually changed the manifest.json to alexapy 1.29.0 Result: Reboot HA. New configuration:
TTS works, switching the "do not disturb" switch does not. |
|
Hi, System details Home Assistant version: 2024.8.3 I changed the values in const.py as in the original post. The integration is working pretty decent, I don't have to reauthenticate every 10 minutes like some days ago. I'm not a programmer or anything, just trying to understand so maybe what I write has no sense or is normal... |
In Summary:
|
OK, all correct! |
|
Hi, I just changed the scan interval to 3600. It's working despite I immediately got this error: 2024-08-27 13:59:48.076 ERROR (MainThread) [homeassistant.config_entries] Error unloading entry ***.*****@.com - amazon.it for notify What does it mean? Before 3600 it was 600 secs and AMP worked for a day. AMP 4.12.11 |
I can't help as I don't use .es... |
This sounds like http2 is not connected... |
Perhaps there are bugs in the dependent Python libraries, and Home Assistant 2024.8.x is using exactly those buggy versions. I have a developer version of HA: 2024.9.x.DEV It's not possible to install it with: I get: The conflict is caused by: But it's possible to overwrite this with, but you shouldn't do it unless you know how to roll back if it breaks: But maybe it's useful to compare the library versions? You can get this list with: |
Oh? I had this issue after an HA upgrade, too. Nonetheless, the integration still worked. I removed the configuration and created a new one, then I could reload without this error. |
Yes, Amazon ceased issuing the command "PUSH_ACTIVITY" via http2push back in June which negated AMP's ability to automatically update LAST_ALEXA via that command. |
You can safely disregard that error as it would seem to be irrelevant. I've witnessed trying to unload a config entry that was never loaded but when I try and troubleshoot the reason then it doesn't happen! If it tries to unload a config entry that isn't loaded it should silently proceed, or maybe issue a warning, but not an error! If I could pinpoint it, I'd try and fix it! |
unfortunately I am not able to give this kind of collaboration. I chose to install OS of HA precisely because I am not able to manage, I'm really sorry 😭 |
|
I have been inside AMP for a while now attempting to dissect it, and comparing it to other "standard" integrations is difficult as nothing else seems to try and do what AMP does, and nothing else is structured quite the way AMP is. |
No worries, thanks! |
|
hello, not clear if someone solve the issues. I have done following change but without good result: To: the reconfiguration pop up appears continuously: just after the HA restart this is the situation: |
Hai risolto? |
Analysis done by ChatGPT: Detailed Chronological Analysis with Focus on Specific Behavior1. Initialization and Start of the Login Process (09:54:08)
2. Failed Registration Attempts (09:54:08 - 09:54:09)
3. Continuing with API Requests Despite Failed Registration (09:54:09 - 09:54:10)
4. Successful Authentication (09:54:10 - 09:54:11)
|
|
I have polling currently disabled but AMP still stops working sometimes, showing After days of testing I'm 99% sure that the issue is not related to I tried TTS and update_last_called 11 times in around 30 minutes right after restarting HA. All worked, not a single issue. Below is what happened during the night, of course nobody was doing anything... AMP then stopped working No interactions made between 2:45 and 13:47... Just posting this as feedback, maybe someone who reads can understand more. |
|
as an update to the above, for now all seems to be working since I posted it. |
Hey Matteo, thanks for your feedback. It's not clear to me what did you do to make it working. |
to be honest I didn't change anything from yesterday and I don't think it will work forever. For sure it is working better. But again... I don't think it's fixed, it's just working longer than usual. This is the same setup I used for days and I think it's just luck that it's now working for around 20 hours... |
Thanks Matteo, i just updated the dependencies via ssh and changed the manifest. Will see if this will stay fixed longer... |
|
I have just updated to the new core 2024.9.0 in order to test if something changed. I found some new errors I didn't see before in core 2024.8.X Error 1: This error is new for me. I didn't remember from the 2024.8.X Error 2: This error occurred under 2024.8.X 10 minutes after the start of HA 2024.9.0 ... the integration fails. I am running ... This time I didn't modify manifest.json file neither the time between scans, I kept the 60s, include devices connected via echo enabled and polling updates. |
|
I am going to update something I did after the first failed test. I restored my 2024.7.4 backup I did before the update to core 2024.9.0. And I decided to start again from the begining. Once the backup was restored ... I downgraded Alexia Media Player from version 4.12.11 to 4.12.7 (I think this was the first release that solved the problem from upgrading from 2024.7.4 to 2024.8.0). After restarting HA I modified the manifest.json file. I modified the file according to this:
Once the file was edited I restarte HA again in order to reload the integration according to this change. After the restart ... AMP 4.12.7 (with modified manifest.json) seems to work as the 4.12.11 under HA Core 2024.7.4. So ... I made a new backup and I decided to upgrade HA to 2024.9.0 again. After the new core has been installed ... I decided to wait ... and after 10 min ... it failed again. So just I found the integration failed (in my case it is easy to check because the temp sensors from the Echos became unavailable) I deleted the pickle file located at .storage folder and I restarted the integration quickly before this file is created again. After HA had restarted I think I found an error at logs, I didn't remember. I thought It didn't work the process, but as I was having dinner I decided to wait until I finished in order to restore my backup to 2024.7.4. But after the dinner I could check the temp sensors did not become unavailable during the last hour. So the integration was still working for more than 10 minutes (without any reload of the integration). I am going to still monitoring the logs until tomorrow in order to see if it doesn't fail. Let's see and ... fingers crossed. |
|
Unfortunately it only worked during 90 min aprox. Then the previous errors 1 & 2 appeared again. After that the integration failed and the reload didn’t solved it. (Error unloading entry Alexa for notify). The sensors become available again but tts was not working. So I restored my 2024.7.4 backup again. :( |
|
A weird thing ... after restoring the backup to core 2024.7.4 ... with AMP 4.12.11 (and no manifest.json edited) ... after a while I could see a couple errors that made me feel scared. Error 1: First time 23:01:03 when HA started after the backup restoration. And second one 23:10:28 ... (the usual 10 minutes when AMP pulls for updates). Error 2: First time 23:01:03 when HA started after the backup restoration. And second one 23:10:28 ... (the usual 10 minutes when AMP pulls for updates). That errors are the errors I saw when the integration fails under 2024.8.X adn 2024.9.0. But this was the first time I saw these errors under 2024.7.4. Every time I restored the backup ... the integration worked without error ... so I was scared and I thought I completely broke the integration, even under 2024.7.4. So, I deleted the pickle file and reload the integration ... and now looks like it is working again under 2024.7.4. |
This is not an issue according to @danielbrunt57 - This should be a warning, not an error.
This was fixed with the latest alexapy 1.29.0 version. |
This is not an AMP issue. These messages were introduced by HA 2024.8 to speed up the HA boot/ initialization process by forcing the programmers to optimize the code. |
|
4.11.2 versions works as expected, with latest HA .9 version. Try to downgrade until the issue is corrected. |
Does version 4.11.2 fix the issue? |
It certainly doesn't correct. It's just that it keeps working, it hasn't stopped working. |
|
Sorry @Frucoboti but I'm having trouble understanding your last 2 comments, maybe it's my fault. In the first post you say:
So you are saying that it works, but that you also should downgrade. Why would I downgrade if it works? In your second post you wrote
Also here you're saying that it doesn't correct the issue, but also that "it keeps working". So I don't understand if it works or not... |
|
Don't worry, I mean: 4.11.2 always has been working, with all HA versions. You can downgrade from any newer version to this, going to HACS, selecting 4.11.2 version.
Better now? :-) |
|
Thanks it's clear now. So, to recap, you're saying the combination HA 2020.9 + AMP 4.11.2 works. |
|
Thank you @Frucoboti for your feedback. Just some more questions about your setup in order to discard if something else is breaking the integration.
|
Mine has been working fine for 48 hours.. I used some TTS and update_last_called is still working at the moment. |
Thats is!
Sorry, your github link don't works for me. But I understand your question. |
|
@Frucoboti for me is not working.
Here my log |
4.12.12 will now remove the |
|
Not sure if related but after installation from HACS I've this popup That opens an amazon login page, I write the user, the password and asks again the user, the password, etc.. Console: |
|
Today I decided to disable polling since AMP lasts maximum 10 minutes after startup. |
|
Has there been any update regarding the problem? Registratore: homeassistant.config_entries Error setting up entry *******************@gmail.com - amazon.it for alexa_media |
|
You will find at least 2-3 semi-working solutions in the last 3-4 weeks of this issue, but to my knowledge and experience none of them are stable enough. We can join with these aproaches, but sooner or later we will be asked to re-authenticate (which is a burden if we need to do these semi-working solutions over and over again). |
|
Luckily I don't use AMP that heavily. I can temporarily do without it, using other devices. But I need it to go back to a stable version so I can use it in the automations I have at the moment. Otherwise it is systematic that when automation is needed AMP is in logoff. I hope that the programmers will be able to find a solution in the next updates, and that the integration will work correctly again like it did a few months ago. |
|
Hi, same for me. error log: |
|
I had this problem from about 3 weeks i have fixed just re-downloading alexa media player from hacs. |
and others
Describe the bug
For those users in the EU region continuing to have issues.
System details
const.pyor HA startup log): 4.12.11pip show alexapyin homeasssistant container or HA startup log): => 1.28.2Debug Logs (alexa_media & alexapy)
Result of successful http2 communication:
What to try
In const.py, try editing these constants as follows:
From:
To:
I'm most interested in whether
MIN_TIME_BETWEEN_FORCED_SCANShas a positive effect or not as the integration only retries 5 times with an increasing backoff in time in between, which may not be enough to prevent Amazon EU from shutting down.It's a shot in the dark since I'm in the NA region (amazon.ca) and do not have your issues...
The text was updated successfully, but these errors were encountered: