@dafish Pleased to continue to receive your counsel - thank you! 

 

Made a significant difference in the Ping Response just by disabling 11b - that was instant 

 

1 single fail on one device in 40 pings over 21 devices 

edit - while writing it's now at 60 pings and just that single fail on that one device, all others at 100% success

 

 

Will have to see if that impacts the overall issue - will likely know in the next few days if that is the issue, or red herring 

But again, confused as to why things suddenly changed (that is not something that I had previously changed to INCLUDE 11b - no need for it) - and the firmware update was just couple of days ago, but should not have changed any settings anyway. 

And I should point out that, although everything is currently showing online in Kasa app, in Alexa, I have two devices that show offline and those two devices in Kasa are not responding as they should (kinda work but not crisp on/off/on when commanded) 

 

 

These are my general wifi settings 

I was wondering about WiFi Agile Multi-band - research suggests a good thing, should I enable this? 

I used a wifi analyzer - my neighbour is on ch 6 for 2.4 and there is nothing else that is close

 

 

This is my Smart Connect Rule 

 

 

 

 

 

Over 450 consecutive pings, I have total of 15 fails over 21 devices; so 15 in 9450; or 0.16% failure or 99.84% success 

 

But after resetting router again earlier after those two flaky connections, it had looked OK for a few hours, but now again have two devices showing offline in Alexa and Kasa 

  @DEcosse 

 

It's entirely possible, if not likely, TP-Link/Alexa have outlying backend issues.

 

It's very good you're seeing more stable pings.   Again, a whole house reset would be good.  I don't trust the switch wifi drivers and believe then can corrupt.

 

Dumping 11b may have helped your router and/or your environment, so good on you for changing that. 

 

As to your settings, I would do this:

 

* Once we know the devices are stable re: ping response I would enable Wifi Agile.  Unless things go south it's good to have.   May not do much in a mesh setting.

* I would move to WPA-3 if the router supported it.

* I don't like mesh (I hard wire each router and use AP mode on the downstream one.   It's faster, more reliable, and I can use different channels for each router), so I'm not steeped in Smart Connect.   That said:

   - 5G Steeping down at 82db seems too high to me.  Not that it matters here (kasa is 2.4 only), but that's pretty weak signal.  I'd step down much sooner.

 

Next:

* Mesh networks, particularly when you get a few "nodes", and they get close to each other, can create problems.  How many do you have?

* For the mesh are you using 5gh #2 for the dedicated backhaul. or..??

   - assuming so, what is signal strength shown at mesh node for link to primary?   aka:  Do you know that backhaul is quiet and stable?

   - one of the problems in mesh networks is folks oft use triband systems to mesh.  Which implies using 5G for the backhaul.  But 5G is relatively distance sensitive and so suddenly our 2.4 units are a little too close to each other or our 5g units are too far away.  I'm wondering if that fits you.

   

* Do you have the equipment and wiring such that you can set up using traditional router plus AP method (hardwired)?

* I presume you selected channel 1 after confirming it was the quietest choice (1, 6, or 11) available?

* Have you checked your switches signal strengths?   Which ones are the weakest, and what are those values?

 

Regardless, if you see stable pings over a few days we can call your network stable and cast an eye at TP_link (and alexa) for the backend stability.  

 

 

 

 

  @dafish I've come to conclusion, in agreement with you, these radios are S#1t! 
I've done a bunch of work tweaking on my Router and got down to where approx 4 devices consistently are the ones failing pings consistently badly.

I tried the whole house power cycle as you suggested but that made no difference (indeed made some of them worse) 

I then went round every one that had even a few ping failures and 'reset' them - the signal strength on each of these improved by 15-20 dBm in each case

(I sure hope this doesn't mean I have to go round and reset every device any time there is a power glitch here!!!) 

Then one of the worst offenders, an HS220 dimmer, was not improving any with multiple reset attempts (even factory reset), so I swapped it out with another I had on hand - instant improvement on the signal and so far it seems like this location is no longer failing a ton of pings. 

The other three bad units all have strong wifi (at least per the app and also my router) 

2 of these are Motion Detect Dimmers ES20M - I have 4 of these altogether, two are rock solid, these other two are high ping failures 

Both of these are very close to routers and measure 34 & 37 dBm signal strength respectively - yet are now my worst ping failures, even though their signal is much better than the other two devices of same type (40 & 58) 

And the last outstanding (careful how you interpret that word!) is an outside dimmer plug KP405 - decent connection again at 42dBm - but this one also has very high ping failures.

Going to report this finding to tech support, especially the one that was resolved by replacement

 

@Riley_S 

 

Among the many suggestions I've made is one that would allow a scheduled restart of smart devices.  I believe this would help stability issues, loss of communications, and etc.  Assuming implemented with a "restart to previous status" capability there is no downside.  Strongly suggest.

 

 

@DEcosse 

 

Progress, even if painful.  Good I guess, right?

 

I don't want to be too hard on Kasa, I bought cheaper units accepting some risk.  Not like I paid Leviton money, right?  On the whole I think the features are pretty impressive.  It's just maddening when stuff's not working.  Still, if I could pay $5 more per device and have better radios I know I'd have paid it.  They are sure oddly quiet about driver updates aren't they?

There is hope:  I've been perfectly stable since.  I've not heard about app updates so I'm still on the beta, but otherwise I believe the general user experience is stable.  I've seen these things (as have you now) can get in a state such that comms is seriously deprecated.  To such a point they can affect other devices too.   I see that as poor drivers, but who knows?

 

Once you get consistent (and you may be here) clean signal levels and ping times I suspect pain will diminish.  Don't forget I've mentioned too many mesh radio's is bad too, right?  I'm supporting 5200 Sq' over three floors on two radio's. 

 

Not sure how restart (which is what I assume you did) is different than a power cycle, so I can't comment.  I've only had to reset a few and haven't had to do it since.  Mind you I some of these are rarely used so I guess I don't know for sure.  (I think I'll check!).  Alexa says they're online, so...

 

FYI, I've had some power events in the area in the interim.  No problem so far.

 

 

Congrats on your progress!

 

 

-d

 

 

 

 

  @dafish I did a complete power cycle at the main breaker (and actually had another one forced on me last night with a power interruption of just a few seconds - but enough to restart everything once again. 

Actually after things looking better yesterday, except for those 4, I went back into the three remaining troublesome ones and did a full factory reset on them - and that resolved the significant ping failures on them. 
Before that power outage everything was looking really stable, with the majority having zero fails and those three particularly were really behaving. 
But what I found is that when things re-power up (after power off), the devices don't necessarily connect to the same router (or who knows what else happens). It seems like they always pick my primary router, which is fine for most devices but I have some devices outside and downstairs (where my second mesh is located).

I only have those two. If I reset the devices in the lower level, they will improve their signal because they connect to that more local router. 

* Actually as I write this I realize that my primary router and cable modem are on a UPS, while the auxiliary is on regular ac outlet - may have something to do with how things reconnect when power comes back on, as the devices may be rebooting quicker than the router (while the primary is staying alive)  

 

Where I am right now:

Everything is in the 1-2% failure range - I'm not sure what is considered 'acceptable' but at least for most part, all the devices are predominantly responding to local requests. 

So those three devices *7, 13 & 48 are at least in line with all the others.

The 'best' performance is *186 which is one of the new Kasa 3-way dimmer which I received in the Beta program; *200 is a Shelly Device and *215 is a smart bulb from another manufacturer. (I actually like the Kasa Bulbs - have a few of them for Table Lamps where it's much simpler than putting smart plugs everywhere)