Failed to deliver message: <EmberStatus.DELIVERY_FAILED: 102>

I’m having the issue too…ran goodnight script via Alexa and dashboard both failed with this error, no problems until most recent update. Has anyone confirmed this is scripts only or are other actions failing?

Same issue here with home assistent yellow. I think it’s an issue with an update of home assistent.

Rebooting home assistent helps. But then after a while there is delay on the network and at some point devices stop responding

I had the same issue. One of my Sonoff TRVZBs stopped responding and now it’s working again after boot. I have ha on rasberry and using Sonoff dongle

1 Like

Also on my side, automations which are failing are working well if i reboot the home-assistant which has skyconnect connected…

This fails sometimes when I set the light information directly from node red, which doesn’t use the scripts

Same problem for me, frequent unreliable zigbee network seems to have started with an update of home assistant in summer 2023.

I use an Intel NUC with Proxmox running HA, ZHA and I have a Skyconnect stick connected via a USB extension cable.

Before summer 2023, everything worked without any issues for years.

This is becoming very frustrating, most of this winter my zigbee TRV’s constantly overheat which will have an impact on my heating bills.

1 Like

exactly the same behavior and setup as on my side. But im using ESXI instead of Proxmox. Also the most annoying failures are the TRV automations… but in my case the house is getting cold, because the automation which increases the temperature is failing at morning…

Hello all, i had this problem few months back and i solved it by change my wifi channel to 11 and zigbee channel to 18 (because lower levels of wifi channel was full of neighbors). After that everythiing works like charm, even much faster. Few weeks back after restore HA one day back from backup i have problem with one device, that i always add but then it is unaviable, and one time i had to restart zigbee. So my next step is again change zigbee channel and repair all network. I think that will be the solution and i will let you knwo if it helps for second time. Just nend to find time when nobody is home :slight_smile:

There is open topic with same problem, seems like there is problem with ZHA, but no solution: EmberStatus.DELIVERY_FAILED: 102 · Issue #103562 · home-assistant/core · GitHub

@Mazoku Have the errors stopped still?

Did not have time to switch channel again. Yes and No.
I had problem with one window/door battery sensor. After replacing battery for second time it started working, so everything was fine. Few days later same sensor but in different room sometimes go unavivable and behave not as should.

I did all updates meanwhile, but now i have one device half broken :slight_smile: (it works when i open window, but then after few hours go unavivable). I have 4 of these sensors, and others works fine for now.

i have to say that i did not see this error since a while / since a few updates. Working well on my side, without any changes to the configuration since the error first appeared.

Hello all,

joining in on this discussion. In the past two weeks or so, the connectivity and performance of most of my motion sensors have been wonky, sometimes extremely bad to the point that I went through over 6 different motion sensors, thinking that they either started breaking due to maybe the cold weather, or just that I’m lately a bit unlucky and all the new ones I purchased are just bad batches. But after reading all the comments here, I’m starting to wonder if it’s either the poor performance of my SkyConnect dongle or an update I did to my HASS.

First off, I am running HASS as a VM on my local Unraid server. The Core version I’m currently on is 2024.1.3, and the OS version is 11.4. My ZHA is off a SkyConnect. I previously had a Sonoff Zigbee Dongle that I migrated over to the SkyConnect.

As mentioned above, about a couple of weeks back, connectivity and performance of my Zigbee motion sensors started becoming unreliable. Most of my Zigbee motion sensors are either Sonoff SNZB-03 or Ouvopo Zigbee motion sensor. Both of them shows up in ZHA as eWelink MS01, and even if the SNZB-03 shows up as that, the manufacturer will still say eWelink. They both function and look the same beyond the physical shell shape. I also have Aqara P1 Motion Sensors in the mix.

3 out of my 4 ewelink motion sensors started acting weirdly in the past 2 weeks. One of SNZB-03 would constantly report that it is out of battery and I had to pull the battery out, put it back in, and it will start working again, but the battery indicator would show that it is very low before it resets back to normal. I swapped that one out with an Ouvopo, but now it would randomly trigger once, just once, every single night, even with the door closed and there are no movements.

The one working ewelink motion sensor was another SNZB-03. I moved that from one room to another location in the house, reset it to get a new name, and all of a sudden, it completely no longer worked. It had an extremely hard time pairing with my ZHA, and when it finally does, the sensors weren’t reporting at all. So after multiple attempts to try and connect and things not working, I chalked it down to the device being broken.

I got two more Ouvopo ewelink sensors, and both constantly had issues. One constantly fired about every hour without any reason whatsoever, and the other one stopped communicating altogether. I haven’t reset it because it’s on a high ceiling and I just don’t want to pull out the ladder right now. I got two more replacements, but I don’t have high hopes for them too.

Even one of my once extremely reliable Aqara P1 motion sensor started acting up. It wouldn’t trigger even if I wave right in front of it but then all of a sudden triggers when I turn my back on it. I finally reset it and now it works like before.

I really don’t think ALL of these motion sensors are defective, there is no way. I have more coming from China via Aliexpress but I don’t want to spend hours testing them, thinking they’re fine, installing them and then finding out hours later that something isn’t right. I really have a strong feeling that either the SkyConnect is the issue here, or HASS needs to update something and fix the reliability issue when connecting to end devices. I’m inclined to switch back to my Sonoff Zigbee Dongle and see if that fixes the issue.

Sorry for the extremely long post. I just needed to rant a bit about how frustrated I’ve been with all this. I intentionally switched nearly all my automations at home over to HASS, telling my wife that it’s much more reliable and whatnot, and here I am running around like a madman trying to fix things that are constantly breaking. I’m lucky my wife finds this whole thing hilarious and doesn’t fuss about lights not turning on etc. UGHH!! :triumph:

Aaaand I forgot the question I wanted to ask all of you LOL!

Some of you mentioned about changing the channel on your ZHA coordinator. I saw this option under network to change the Channel directly. Can I do it here? I did a bit of googling on how to change the Zigbee channel, and none of them mentioned about doing it here. It’s always some long convoluted steps that always end with “this will break your network and you have to reset EVERYTHING!” lol

I changed my channel on the screen you indicated. You might first want to “Download diagnostics” from your main Zigbee device (click three dots to the right and download diagnostics there). Look at the file to see what channel has the least utilization. I’m currently on channel 16 and while I still get the occasional DELIVERY FAILED: 102 error, it’s not too often that it’s a problem. Good luck.

1 Like

Hi @ecchodun, thanks for your reply. I downloaded the diagnostic from my SkyConnect (Zigbee Coordinator under Device), and looking through the content, I can’t seem to find anything that refers to different channel utilization. I most likely don’t know what to look for exactly; still new to this whole thing :smiley:

Mind pointing out where in the text file I should be looking at?

It will look like this -

“energy_scan”:
“11”: 28.30261646762903,
“12”: 25.74050169409602,
“13”: 43.057636198227904,
“14”: 46.26944564832987,
“15”: 59.15797905332195,
“16”: 46.26944564832987,
“17”: 68.14622793558128,
“18”: 98.43344238842926,
“19”: 99.363464769115,
“20”: 99.06269548719737,
“21”: 98.93395819824867,
“22”: 75.96022321405563,
“23”: 65.26028270288712,
“24”: 9.713248103580147,
“25”: 17.086630587133605,
“26”: 59.15797905332195

Those are channels and the second number is utilization. Lesser is better.

1 Like

Oh I downloaded the wrong diagnostics LOL! I was staring at the diagnostics of the coordinator itself, but instead, I was supposed to download the diagnostic for the whole Zigbee network under Integration Entries. Now I see it.

Mine looks like this -

“energy_scan”: {
“11”: 91.05606689948522,
“12”: 97.7033852118351,
“13”: 97.39286236923465,
“14”: 89.93931580241996,
“15”: 97.97769123383605,
“16”: 97.39286236923465,
“17”: 98.62178092672917,
“18”: 97.97769123383605,
“19”: 97.39286236923465,
“20”: 52.75969252664325,
“21”: 52.75969252664325,
“22”: 78.25348754651363,
“23”: 82.35373987514762,
“24”: 73.50699819621309,
“25”: 92.95959997754716,
“26”: 98.21983128611214

So if lesser is better, then should I go with channel 20 or 21? I’m currently on Channel 11.

That is correct. It may take a bit for things to switch over completely (about an hour). Many of those numbers do look high so you could have some wifi interference. You might need to switch your wifi router channel as well.

1 Like

Awesome! Thank you so much for your help. Crossing my fingers in hopes that the new channel will help fix some of the issues I’m having. Will update here on how things goes.

Hi @ecchodun, unfortunately, and surprisingly, changing the channel to 20 seems to end up being worse for me. Not sure if this is related or not, but now all of my signal strength between devices seems to have dropped lower. Their LQI readings are quite lower than before. Looking at my network visualization, previously I had a couple of green lines and mostly yellow with a couple of reds. Now it’s all reds with two yellows.

I am somewhat convinced that the root of my problem might be the signal strength that the SkyConnect provides and the reliance on the end devices having to connect to the routers while the routers themselves don’t have a good connection to the SkyConnect to begin with. I need to re-verify this, but I believe with the Sonoff Dongle, the overall interconnectivity of the devices is much better, thus resulting of course in a more stable and reliable environment.

Is there a way to “hack” and maybe add some kind of antenna to the SkyConnect to improve its signal range? If not, then I might need to move back to the Sonoff Dongle.