60.2 (Beta) Firmware API problems

[scharrrfi]

Hi,

I have updated to the wallbox beta firmware 60.2, because of the 15-minutes tibber issues.

Now the API connection of my Home Assistant to the wallbox works only for 47-72 hours. After this, it disconnects and my HA logs

• unexpected: [Errno 104] Connection reset by peer
• unexpected: Server disconnected

The only way to reconnect at the moment is to reboot the wallbox. I am not sure, if this is a firmware only bug or something has changed in the API what causes the crash. The wallbox can control normally over the app when the error occurs.

Thank your for investigating this.

[marq24]

I am sorry for the late reply - I just update my WB to 60.2 - then let's wait additional 48h in order to see, if this is a general issue (could be that the new Firmware will auto close the HTTP Session after a certain time)...

[marq24]

also what would already be helpful for me is to get the full HA error log message (there might be a trace?)

[Olgrov]

I got the same issue. Here are some logs:

Dieser Fehler stammt von einer benutzerdefinierten Integration

Logger: custom_components.goecharger_api2
Quelle: custom_components/goecharger_api2/config_flow.py:235
Integration: go-e APIv2 Connect (Dokumentation, Probleme)
Erstmals aufgetreten: 08:08:07 (3 Vorkommnisse)
Zuletzt protokolliert: 08:11:42

Exception while test credentials: [Errno 104] Connection reset by peer

Dieser Fehler stammt von einer benutzerdefinierten Integration

Logger: custom_components.goecharger_api2
Quelle: custom_components/goecharger_api2/init.py:272
Integration: go-e APIv2 Connect (Dokumentation, Probleme)
Erstmals aufgetreten: 18. Dezember 2025 um 22:25:46 (447 Vorkommnisse)
Zuletzt protokolliert: 08:11:50

unexpected: Server disconnected
unexpected: [Errno 104] Connection reset by peer

Dieser Fehler stammt von einer benutzerdefinierten Integration

Logger: custom_components.goecharger_api2
Quelle: helpers/update_coordinator.py:461
Integration: go-e APIv2 Connect (Dokumentation, Probleme)
Erstmals aufgetreten: 18. Dezember 2025 um 22:25:46 (447 Vorkommnisse)
Zuletzt protokolliert: 08:11:50

Error fetching goecharger_api2 data:

[marq24]

the main issue right now for me is, that my HA does not run for 48h (without a restart cause of an Integration update) - so I really can't tell, if this is a general issue...

What I would like to know @Olgrov - are you using the LOCAL Version (so accessing the Wallbox directly) - or are you using the Cloud variant where the data will be fetched via https://[your-serial-here].api.v3.go-e.io/ ?!

[Olgrov]

I am using the local HTTP-API v2

[marq24]

thanks for the clarification... this one is really strange:

Dieser Fehler stammt von einer benutzerdefinierten Integration

Logger: custom_components.goecharger_api2
Quelle: custom_components/goecharger_api2/config_flow.py:235
Integration: go-e APIv2 Connect (Dokumentation, Probleme)
Erstmals aufgetreten: 08:08:07 (3 Vorkommnisse)
Zuletzt protokolliert: 08:11:42

Exception while test credentials: [Errno 104] Connection reset by peer

Since this is during the setup / configuration phase - and even there you get instantly a Connection reset by peer while the session is just created fresh.

I don't have a good explanation for this - and would kindly ask, if you can double check your network setup?!

[Olgrov]

I just did. I've figured out that my goe-App was connected via Cloud only, even though I was using the same WiFi. Hence, I would assume that the local connection to the wallbox is buggy in general (because also the app could only connect via cloud). After a restart of the wallbox, the app could connect again via local access. Now this integration can access the wallbox again via local api.
Seems to me a bug of the latest firmware, I guess. I will wait a few days and report if the error will re-occur.

[scharrrfi]

For me it seems now to be stable for more than 72 hours. I will post here if the error occurs again.

[dns13]

At least for me, the last extension update has changed nothing. But in my opinion it can't. If it fails, it seems that the complete local api is crashing and this is only fixed by a reboot of the Wallbox. I was hoping for a fix in 60.3 beta but no, still happening.
It seems to be related by using the app for me - it can run for days without issues, but if it fails it happens directly after doing stuff in the app, but don't seem to be a specific thing.
Hopefully this will be fixed until the stable release.

[marq24]

Thanks for your update... Since I am not using the goe App at all, I guess we must wait for an updated Firmware. (I had also the hope that the 60.3 will solve it)... So lets keep this issue open - even if we both agree that it seams to be an issue of the Device and not of the Integration.

[dns13]

I just rolled back to 59.4 and quickly ran into the same issue, again. If the charger is in the failed state, also the firmware download is not working before a restart.

I did not have this issue with 59.4 before updating to the 60 beta. That may mean:

• Updating to 60 beta breaks something in the firmware that is not rolled back with downgrading
• They fucked up something in their cloud service when updating to new 15 minute prices that also affects older Firmware
• Maybe a regression in the extension when adapting to new api keys in 60 Beta

I can reproduce it reliably just by setting various registers a few times back and forth. Tried with awp, lmo and trx but I suspect it's on any register.

Update: I cannot trigger it by simple GET requests to like http://10.0.14.235/api/set?awp=10
Update2: Just tried to see whats going on with tcpdump, but surely I cannot reproduce it now. :(

[fabianwimberger]

The downgrade not fixing it suprises me a bit. The line I though may be the reason is "Brute Force login protection in local network" of 60.0.

I will check if 60.3 maybe works fine (just upgraded) and if it happens again, I go back to 59.4 seeing if it also happens there now.

PS: For me the disconnect happened exactly when i start to charge the car (via nfc card). Was pretty reproducible.

[dns13]

Yes it feels like it's related to whether a car is connected/charging or not. Perfectly stable since my last comment - but there was no car connected since then. But may be biased as I only actively do stuff if a car is connected...

[marq24]

Have in mind, that when the wallbox is not connected with a vehicle, the integration is requesting way less data via the API - see also here https://github.com/marq24/ha-goecharger-api2?tab=readme-ov-file#hibernation-mode--good-to-know

And when this is a (false positiv) DoS protection issue, then what is the current update interval you have configured? - I use the default 5 sec here.

[dns13]

Good point, that would explain it. I'm using 30s update interval. Maybe it's related on how the requests are bunched together, so not directly the update interval but the burst when requesting different datasets? But what I've seen on the tcpdump it's far from spammy - their brute force protection must be very prohibitive.