[BUG] HA becomes unresponsive due to Omnibattery integration V1.5.0b2

[Stino87]

Affected area

Other reproducible behavior

Omnibattery version

1.5.0b2

Home Assistant version

2026.9.2

Installation details

Battery brand and exact model: Zendure 2400AC+
Battery firmware version: V2.0.2 (latest)
Connection method or adapter: wifi
Number of batteries: 1 (2 packs 5kWh total)
Consumption sensor and update interval: P1 every 1-4s
Installation type (single-phase or three-phase): single fase (NL)

What happened?

HA becomes unresponsive when enabling Omnibattery integration (few minutes to load between pages or just giving errors).

Steps to reproduce

Just enabling Omnibattery integration. I have reproduced this error 3-4 times now. Sometimes it works for a few days/hours, and sometimes it crashes immediately.
I did a redownload to V1.4.0. It doesn't seem to fix the problem.

Expected behavior

not crash HA.

Relevant Home Assistant logs

No relevant log entries.

Diagnostics file

I managed to disable Omnibattery. Start debug. enable Omnibattery. wait 2 minutes and disable it to get the debug file.

home-assistant_omnibattery_2026-09-18T08-14-44.299Z_public.log

Additional context

No response

Confirmations

• I am using the latest available Omnibattery release and restarted Home Assistant after updating.
• I searched the documentation, existing issues, and existing discussions.
• I can reproduce this unexpected behavior; this is not a request for installation or configuration help.
• I removed passwords, tokens, API keys, complete serial numbers, IP addresses, and other sensitive information.

[ffunes]

Thanks for the log. It doesn't point at Omnibattery, though. Omnibattery only loads at 10:11. In the hour before that (09:06 → 10:11), HA was already struggling:

• go2rtc was crash-looping about every 3 minutes ("didn't terminate gracefully. Killing it")
• Supervisor API timeouts
• "Something is blocking Home Assistant from wrapping up the start up phase"
• zendure_ha hitting TimeoutError during httpGet every minute

You also have three integrations talking to the same 2400 AC+: zendure_ha, battery_smartflow_ai and Omnibattery. The device's local HTTP API was already timing out with zendure_ha alone. Once Omnibattery joins, its polls go from 0.06 s to 6–9 s. And they will fight over the setpoints. Please only let one integration control the battery.

Could you try this?

1. Disable zendure_ha and battery_smartflow_ai (or at least remove the Zendure device from them) and run Omnibattery on its own. Does HA stay responsive?
2. The other way round: keep Omnibattery disabled for a day. Do the go2rtc restarts and slowness still happen?
3. Tell me your host type and the CPU/RAM usage from Settings → System → Hardware.

If HA is still slow with only Omnibattery running, install the Profiler integration, run profiler.start for 60 s while it's slow, and attach the output. That will show what is blocking the event loop.

[Stino87]

Thanks for the quick reply. I'm gonna check this out and let you know if the problem still occurs.