ebus connection stops working after some uptime

[amichelic]

Have an ebus adapter v5 with ESP32-C3, rev 4. When running it with firmware 20250615, I noticed that it stops reporting ebus data after a while, and the log is filled up with these alternating messages:

20:38:45 Warning ebus: errors: overrun
20:38:45 Info host: history: @-56999 >ec81 2001 >ec81 3998 >ec81 3999 >ec81 2000 >ec81 3999 >ec81 4000 >ec81 1000 >ec81 3999 >ec81 3999 >ec81 2000 >ec81 1000 >ec81 3998 >ec81 4000 >ec81 1000 >ec81 4000 >ec81 3999 >ec81 999 >ec81 1000 >ec81 3999 >ec81
20:38:50 Warning ebus: errors: overrun
20:38:50 Info host: history: @-55999 >ec81 3999 >ec81 2000 >ec81 3999 >ec81 4000 >ec81 1000 >ec81 3999 >ec81 3999 >ec81 2000 >ec81 1000 >ec81 3998 >ec81 4000 >ec81 1000 >ec81 4000 >ec81 3999 >ec81 999 >ec81 1000 >ec81 3999 >ec81 4000 >ec81 1999 >ec81

(and so on, a new pair of messages every 5 seconds). I have never yet seen to beginning of these messages - after a reboot, it works fine for several hours or sometimes even 2 days; but when the problems start, the log quickly fills up, overwriting all old messages.

Firmware 20250415 seems to run fine, most of the time there's only the hourly message about the updated NTP time. Data is reported to ebusd/Home Assistant continuously. I noticed however that it seems to reboot after a few days; have followed the uptime over 2 weeks, and I never saw more than 2-3 days.

What could be going wrong here? Is there any possibility of logging to an external server?

[john30]

please check again with the current version, which also adds the option to send logs to syslog

[amichelic]

Thank you! I've installed the updated version and set up logging to an external host. Will update this issue with the results in a few days.

[amichelic]

The option to log to syslog is great, thank you!

Turns out it stops reporting data when the WiFi system it's doing it's nightly frequency sweep at 03:00 (my APs don't have a dedicated radio for this, so they interrupt WiFi and clients would need to connect to a different AP). First night it managed to keep the ebus connection, second night it broke:

2025-11-04T03:01:35.312373+01:00 ebus-98198c cmd_wifi: sta connect, set fix ip
2025-11-04T03:01:35.312373+01:00 ebus-98198c repl_event_task: sta event: 1
2025-11-04T03:01:41.308889+01:00 ebus-98198c wait_host_tcp: accepted: 192.168.10.4
2025-11-04T03:01:41.308889+01:00 ebus-98198c loop: host reset
2025-11-04T03:41:24.823531+01:00 ebus-98198c sntp: got NTP time: 2025-11-04 02:41:24
…
2025-11-05T02:48:56.288606+01:00 ebus-98198c sntp: got NTP time: 2025-11-05 01:48:55
2025-11-05T03:02:36.406127+01:00 ebus-98198c cmd_wifi: sta connect, set fix ip
2025-11-05T03:02:36.407902+01:00 ebus-98198c repl_event_task: sta event: 1
2025-11-05T03:02:38.398063+01:00 ebus-98198c ebus: errors: overrun
2025-11-05T03:02:38.400558+01:00 ebus-98198c ebus: history: @-7896 >03 5 >00 30 >16 41 >00 31 >8a 74 >00 36 >aa 1577 >15 36 >b5 36 >24 5 >06 5 >02 37 >00 33 >03 35 >00 36 >12 6 >00 6 >4b 63 >00 6 >aa
2025-11-05T03:02:38.414862+01:00 ebus-98198c host: history: @-5740 >c6aa 43 >c6aa 44 >c6aa 43 >c6aa 43 >c6aa 43 >c6aa 43 >c6aa 43 >c6aa 43 >c6aa 43 >c6aa 43 >c6aa 43 >c6aa 43 >c6aa 43 >c6aa 43 >c6aa 43 >c6aa 43 >c6aa 17 <c6aa+126 1 <c6aa+126 0 C80
2025-11-05T03:02:42.903253+01:00 ebus-98198c wait_host_tcp: accepted: 192.168.10.4
2025-11-05T03:02:42.904910+01:00 ebus-98198c loop: host reset
2025-11-05T03:48:56.510824+01:00 ebus-98198c sntp: got NTP time: 2025-11-05 02:48:56

I had then tried a few days with the frequency sweep disabled, no communication breaks in this time. I did then re-enable the frequency sweep and also the watch dog (set to 120s, restart eBUS I/O) and last night it bailed out again. In the logs I see a surge of the same errors as above, some messages related to TCP connections and at 03:03:55 the watchdog reacting and restarting eBUS, after which it continued to work fine (can provide full logs, skipped them here as they were a bit verbose).

Tried to use a USR-ES1 module that I had lying around, but didn't manage to get it detected, so maybe it's faulty. Now waiting for a new one. For the meantime it looks like the watchdog can at least mitigate the problems.

[amichelic]

Got a new USR-ES1 module, and this one seems to be working fine. Not a single glitch since I switched over to a cabled connection. Thanks again for the syslog option, very much appreciated!