Using a TP-Link Archer C7v5 here. Don’t experience the problem too often, but often enough to be annoying when automations don’t work due to device being “unavailable”.
1 Like
Using the Bell Home Hub 3000
I have set mine up now with node-red and virtual entities. This allowed me to get rid of the HA integration and still be able to use the switches in automations. An added bonus is that I can change the polling time for the switches that matter.
1 Like
@silverl I have 5 HS300s and am running a UDM Pro + 3 active APs - primary is a new U6-Lite, and have an in-wall AC HD (basement) and AC Mesh (garage) for coverage (lower power, different channel). I still have regular ‘setup / updating tplink took over x seconds’ errors in my logs - I suspected wifi was the culprit (low rssi and/or split channels) so experimented with disabling the secondary APs vs running 3 APs on different channels / power to fine-tune coverage, but didn’t see any change in error rates. I also added a fw rule to restrict the HS300s from the internet - still receive the errors, but the devices don’t seem to actually go ‘unavailable’ (subjective / not quantified) = automations seem to work consistently. Maybe a recent HA update has added some resiliency to the TPLink error handling as well (there have been a few over the last couple weeks).
edit also running 2x pi-holes for DNS (primary/secondary), but my UDM handles DHCP
Here to report what everyone else reports. This integration is unreliable. Here is my setup and observations:
- around 20 Kasa US-style devices (in Canada): 1way, 3way, dimmers, simple wall plugs, and an outdoor double plug
- all have static IP for a couple of months; auto-discovery false
- 5 node wifi mesh everywhere (Netgear Orbi RBx50); 2.4GHz signal excellent everywhere; devices ultimately arrange themselves to connect to the correct closest satellite; going unavailable not correlated to which satellite connected or to signal strength
- lots of log messages that took too long to respond, but those generally don’t correlate to going unavailable
- various devices of all types randomly go unavailable; usually, they fix themselves within 24 hours, but I have one now (HS210) that seems to have permanently gone unavailable
- restarting HA doesn’t fix, rebooting host doesn’t fix
- when unavailable, 100% accessible over Kasa app and over Alexa Kasa skill; also 100% accessible using Node Red TP-Link node (Node Red addon in HA, so same host and network interface as HA)
As was just suggested, several years is well past time for the reliability of this HA integration to have been fixed. Since it’s not, it seems it is abandoned. While it is a bunch of work, I am going to convert it all to Node Red-based entities and abandon this integration as well.
1 Like
Since I don’t see it mentioned in this topic, I wanted to add that I have worked around this flakey “TP-Link Kasa Smart” integration by enabling the TP-Link integration within SmartThings. The SmartThings Home Assistant integration then makes these switches available in Home Assistant.
Even if you weren’t using SmartThings before, installing the app, adding TP-Link then installing the SmartTings Home Assistant integration should make these switches show reliably within Home Assistant.
1 Like
Does that use local Kasa device API or cloud? If cloud, well …
I assume it uses TP-Link Cloud. Merely a stopgap until there is local API support I suppose.
For me although local is preferable my main problem was unavailable switches in the TP-Link integration. Loading them via SmartThings seems to solve that.
1 Like
The issue is that I think we’re all on borrowed time on TP-Link Kasa with respect to ability to maintain future local control. TP-Link is showing all the signs of working on preventing local control and so I think we’re in a race now. At least I am: need to convert all Kasa to 100% local (with existing automation and Alexa functionality) and block all firmware and other cloud communications for all Kasa devices … before they push a crippling update.
2 Likes
I just came here to say that I too was having a similar issue.
my TP Link switches have actually been rock solid. I’ve had them since 2018 and this is the first time I’ve experienced this problem.
- Restarting HA did not fix my issue.
- Reloading the integration did not fix my issue.
- Deleting and reinstalling did not fix it either.
I have 14 of these switches and dimmers and I thought that they were all showing as unavailable, however, I noticed that 3 of them were in fact online. Those devices are installed upstairs and are on a different AP. So for fun, I restarted the AP on the main floor and voila, the other 11 switches are back online.
Here’s the odd part… not 1 minute later, the original 3 connected upstairs are now showing as unavaialble. I restart HA, reload the integration… nothing.
I reboot the AP upstairs… and now we’re back in business.
Super weird.
how do your 2 access points work together? if there is handover to each other and no/forced disassociation happening, the devices might lose the connection without knowing it.
I had to force my tp switches to a specific ssid on one access point to keep more stability. The multi AP one SSID wasnt really helping with these things
To add my experience with these plugs. I have a BT Smart Hub 2 with a number of the black discs. Until 46 days ago I was having all sorts of problems with not only the TP-Link devices dropping out, but other devices abandoning their nearby disc in favour of the much weaker main hub (specifically some bird box cameras).
Then, 46 days ago, BT updated the firmware on the Smart Hub 2, since when not a single problem with any device. They all stay connected to their nearest/strongest AP, whilst my phone switches seamlessly between APs as a wander around the house.
So it’s worth checking, especially if you have multiple APs, that your firmware is fully up-to-date and the forums for your device are not reporting strange glitches with connectivity. I know that the BT/Sonos forum are reporting similar drop-out problems.
Both my APs are connected via ethernet to my router.
I have 3 SSIDs, each floor has it’s own 2.4GHz SSID and my 5GHZ roams around.
This forces all my IoT devices to a specific AP.
Same problem. I bought a 4 pack of EP10 smart plugs. Initially, they all worked. Then they all went offline at the same time for about 12 hours, and only 2 of them reconnected to my router. All 4 have TX, only 2 have RX. What a joke. I bought these to replace X10 modules. Turns out tech from the '70s is more reliable.
Edit: They all work now and I’m happy again. I manually assigned IP addresses instead of letting DHCP do it and I also bound them to the closest access point and all seems to be good.
All my smart home plugs are Kasa and I have a bulb in my bedroom as my kasa. 90% of the time they are great and the power strips in particular are very useful but maybe once a month they have a fit and I need to reset everything to get them on HA again. This morning I noticed they were all gone again but nothing I usually do has worked to restore them this time. I even factory rest one my power strips and then tried deleting the integration and starting from scratch but the HA integration won’t even register any of my TP link devices anymore. I set static IP’s for all of them and then manually added them to the config but with no success. Looks like it might be time to migrate my set up to a new provider as the TP link integration has just been too unreliable over the past couple of months. Lesson learnt. Best to diversify as much as possible when it comes to device providers and wireless protocols
Update: After adding all the static IP’s to the manual config and wait about an hour the devices just suddenly started showing one by one. It happened after I went to look at the settings on the device through Alexa, I’m not sure if that’s related. Maybe that ‘kickstarted’ it back into action or something. Regardless I think I’m gonna be keeping my eyes peeled for alternatives
I was having same unavailable issues randomly on 20+ plugs(turns on/off pinball machines). I also happen to use hubitat C7 hub only for zwave and have the hubitat hub connected to HA via MQTT so i added the tp-link/Kasa outlets to hubitat instead of HA and seems to have fixed my issues. Not a great fix but it does seem to work now.
So maybe we need to collect data on model and hardware version, mine do also randomly go in/out. I previously had a bunch of HS200 V1 and HS105 V1 and they worked perfectly with HA and the app.
Its been a year or so and I had to swap out the HS200, at the time I got HS200 V3. The SD card also died at this time and I never got around to reinstalling HA until recently.
I had inconstancies with the app as well thinking it was my TPlink EAP APs, but have tested each AP individually and with roaming. Making sure no other devices are causing issues with the kasa devices.
The conclusion I have come up with is that V3 hardware as reported in the app have issues, this is the case with HS200 and HS105. I also have HS103 V5 and HS105 V1 and they are rock stable in the app and HA.
Reliability in the app seems allot worse now as well after the most recent firmware update for both HS200V3 1.1.5 and HS105V3 1.1.4.
Was this the fix? Add retries for tplink discovery (#52015) · home-assistant/core@8d8af60 (github.com)
My smart lights (KL125’s) are still disconnecting from Home Assistant, but not the Kasa app
oh man. one of those newbies who dropped a bunch of cash on kp115s thinking they’d integrate awesomely into HA and now they keep going “unavailable”.
2 Likes