Wemo Devices Lost Connection - Reconnected Every 5 minutes

I’m having some strange issues with my wemo devices every 5 minutes or so they’ll lose connection then 1 sec later reconnect? This is only effecting my wemo lights / switches, 8 of them. The switches seem to be effected more than the light bulbs. All my other switches / lights seem fine (mqtt, lifx, tp-link, zwave).



My setup:

Things I’ve tried so far:

  • Upgrading to 92.2
  • Upgrading pywemo from 0.4.34 to 0.4.39
  • Messing with settings on my wireless router

Fixed my issue by altering the wemo code and placing it in the custom components. I altered the line

with async_timeout.timeout(5):
with async_timeout.timeout(30):

Now my timed automation’s can fire. Before the timer would get reset every 5 mins when the device would go unavailable.

I’m still seeing issues in .100.3. I wonder why the timeouts are set so short?

I’m having this same issue.
Lot’s of errors like this:
Error communicating with Wemo Mini, retry 0
Lost connection to Wemo Mini
Unable to re-probe wemo at 192.168.X.XX

Any other recommendations? Currently on 0.102.1.

I came across this thread because I started having Wemo reconnects every 5 minutes with a history graph that looked very much like the picture in the first post. Rather than increase the 5 second timeout, it occurred to me that 5 seconds is an awfully long time for a device to respond - maybe my network could use a little TLC. The signal strengths were all good, and no channel interference according to spectrum scans, so what could be slowing things down so much? In my case it was simply a combination of too many devices and one problematic laptop that I was using as an HA dashboard - turns out the old Intel built-in 2200BG wifi on my Sony Vaio doesn’t behave too well with Debian drivers. An external usb wifi adaptor fixed that. I also shifted a dozen-or-so devices from the main network to a spare router running in client mode, which effectively created a subnet that could quarantine all those UDP broadcast packets. No more HA timeouts now!

@jambalaya I think you’re probably correct with your assumption. I believe the code change for me fixed the symptom not the underlying problem. My wifi network was constantly being thrashed with around 60 devices connected constantly. I’ve recently updated my home network replacing my ddwrt routers with Ubiquiti equipment, following this video. Since I’ve updated my wifi setup I have not seen this issue using the stock wemo code.

Thanks for the Ubiquiti equipment pointer - good to know there is a wifi solution that can cope with 60+ devices, although at quite a cost premium it seems! For now I will stick with spare routers on separate subnets to split things up a bit. I still have one Wemo timeout every few days which looks like it is related to something Home Assistant itself might be doing in the wee hours of the morning. The only clue is that the Raspberry Pi memory usage jumps by 3% at the same time - not sure whether that is the cause or the effect:

While I am looking at history graphs - I really wish the timescale of graphs with vs without unit-of-measure would fully align to make it easier to correlate events in time…