Maybe once a big chunk of processing is done locally for a majority of users it will free up cloud processing resources. That said, SmartThings should have plenty of bandwidth and other resources. If they are indeed overloaded that would be a shame. With AWS and the many like services there would be no excuse for that.
2 Likes
I guess I think the issue isnât so much with the simple things, but rather the more complex, complete solutions that people have creatively developed. I see these solutions as being more burdensome to the cloud. I hope youâre right.
I see plain-vanilla stuff fail all the time. Just a simple hello home action, nothing custom at all.
The hub also goes off-line without any pre-announcements, and changes get pushed out without pre-announcements. That can have a big impact on someone like me.
The hub went off-line for about 20 minutes around 7:10 in the morning a couple days ago, part of all the various platform changes and fixes, but no warning. And no way to postpone it. As it happened, I was alone in the house except for my dog, because my morning aide didnât come that early and my housemates had already left. I would have much preferred to postpone it until someone else would be there.
But I can certainly verify that itâs not just the very sophisticated custom apps like smartalarms that fail. I do as many things as possible with non-custom options, and I still see fails every week.
Obviously I donât know yet whether that will change with V2.
2 Likes
I see plain-vanilla stuff fail all the time. Just a simple hello home action, nothing custom at all.
But I can certainly verify that itâs not just the very sophisticated custom apps like smartalarms that fail. I do as many things as possible with non-custom options, and I still see fails every week.
That differs from my experience. All of my basic stuff works well. The worst thing Iâve seen is a noticeable lag here or there. Maybe Iâve just been extremely lucky.
Curious, but do you only use stock apps and device types? I ask, because my theory (eluded to previously), is that if you have any custom SmartApps or Device Types in the IDE, those are the ones that can fail. I have seen stock apps fail when they are tied to custom device types (and vice versa). But from my point of view, that now makes it custom. I have NEVER seen the pure stock ones fail - EVER. Please keep in mind, my failures have all been associated with the 20 app second error (and those are the failures that I was talking about above). Other than that, my system has been great (178 devices - many custom device types and numerous SmartApps).
By the way, for example, if Hello Home is used trigger your thermostats and they are Ecobee 3âs (like I use), my understanding all along, was that if any app called any device that was tied to the cloud, it would become cloud bound as well.
changes get pushed out without pre-announcements
I agree wholeheartedly, when changes are pushed out without notice, I experience failures and strange behavior. I suspect that even if I were notified, the problems would still exist - I would just have the heads up (which would be appreciated). Sometimes I need to change something, but most times the issues work themselves out as ST fixes the newly introduced bugs. I find this extremely frustrating. However I categorize these issues differently than the ones previously discussed.
Which is all well and good but youâre now dependent on both manufacturers to maintain compatibility. ST is at the mercy of Belkin and trusting that if they change anything, theyâll let ST know, sadly, thatâs generally not the case.
Are you signed up for alerts from status.smartthings.com? The other morning I received alerts that there were issues.
I donât trust the cross platform integration. Belkin can release an update that breaks it and so could SmartThings. I am going to replace the two WeMo switches I have with GE switches. All my GE switches have been rock solid.
So far for me too⊠that being said the same applies to all manufacturers/any devices that ST does not have control over the firmware.
Sorry, but your theory is incorrect. I have custom SmartApps that run forever with no glitches at all. The cloud has glitches as you say, but all apps are punished by those, not just custom.
Okay, really busy week for last week⊠I must have missed something. Whatâs the really disappointing stuff you learned about V2?
I have custom apps that have never failed too. Itâs just that the only ones that I have had fail are custom. Theory still plausible.
As @JDRoberts says, I have built-in apps fail every bit as often as custom. Hello Home phrases that donât run, built-in âTurn on when motionâ missing motion-active events, etc. etc. It happens all of the time, in a relative sense. Itâs actually rare for any of my custom apps to fail, beyond lost events that vanish in the cloud.
1 Like
Many people have had pure vanilla stock setups fail. Just a few examples off the top of my head:
-
sunset triggers failed for many people at various times, even something as simple as switching to night mode at sunset. Smartthings acknowledged this publicly a number of times.
-
there was an issue for two or three weeks where minimotes using the standard stock device handler failed. Again, acknowledged by smartthings at the time.
-
The stock presence sensors for both phones and zigbee fob failed at times for many people. Also acknowledged by SmartThings.
-
I have a SmartThings branded motion sensor used as a touchless switch with a regular hello home action. It gets used about three times a day, and fails about four times a month. (Itâs very noticeable when it fails.) It is less than a year old. Thatâs a failure rate of over 3%, very high for a standard stock action.
-
I have a current problem wear a zigbee dimmer switch is not correctly identified to echo. Standard switch sold by smartthings in their lighting kit. Stock device handler. Only associated with stock apps.
I use very few custom apps or device handlers, mostly because itâs difficult for me to install them. And I still get failures at least once a week.
The thing is, cloud âdisruptionsâ occur frequently. Many of these never rise to the level of something on STâs status page. You can see it when a motion activated light takes 5 or 10 seconds to turn on, instead of right away. You can see it when motion events are lost (the appâs subscription fails somehow). And to be fair, there are times that the internet introduces disruptions that one could blame on the cloud (however, whenever Iâve investigated lags by studying logs, the finger points to the cloud, not the internet). Sometimes the rates of things flaking out increases, as if the ST cloud machine is in some difficult state. Usually, the rate of flaking out is low, thankfully, or none of us would stick around.
The hope for V2 is that we could get simple apps to run locally instead of in the cloud, which would remove these cloud issues from STâs performance. That feature is off in the futureâŠ
2 Likes
Good pointsâŠ
I agree. I was hoping that any cloud-free apps could run local eliminating the cloud penalty, whether it be performance or throttling. I just find the new revelation frustrating.
3 Likes
Sorry @japple , but I know nothing about Staples connect stability Afaik ( which is not very far ) it was mostly proprietary zigbee protocol that did not like to play with others.
Much like the system ( name escapes me) Loweâs branded. As long as you purchased all your sensors from Loweâs you were âokayâ. Donât try buying a GE or Cree bulb at HD and pair it with your Loweâs hub. Although the Loweâs smoke detectors were just rebranded FA and worked with almost any hub.
1 Like
Disappointing for me (and apparently others) but perhaps not a problem for you. I think ST raised expectations about the capability of local execution with V2 hub, but it appears those expectations (at least mine) will not be met, initially at least. A search of my posts in the last couple of days, and there havenât been that many, will give you some idea. (sorry, not sure how to directly link to a previous post) Those of @JDRoberts @geko and many others will expose additional problems, related more specifically to the new approval process required for apps to run locally. Additionally, there are just a lot of reports about inconsistent operations (nothing new here), broken apps, etc. ST has a lot on their plate right now, and I understand there is no perfect system, but I donât need the aggravation of being on the bleeding edge of V2, when even V1 isnât totally stable. Iâm sure they will eventually get everything straightened out, but I would prefer to wait until they are able to deliver a stable product with a wide variety of apps that are capable of true local execution.
This is what I am thinking.
That said considering on a long drive today all of this, I am moving a lot of my custom apps from the community into things like Smart Lighting, as it handles much of what I need. That will go local. A lot of my PEQ, ST stuff, etc is all their device types/stock smart apps, so those will all go local. A good part of my stuff can go local because I did focus on building out with core stuff. I do have edge things, but if there are a lot like me that moves 80% of their processing (and sunrise/sunset actions) to local, this may free up cloud resources.
That said, they have a dual effort of insuring updates work cloud and local in âfirmwareâ so will random errors go away? I doubt it completely, but I have also seen things that are just due to the actual devices going funky or other anomalies. So it takes some doing to assign the blame. But I do, I do. I re-architected my setup maybe 6 months ago, stopped going ape with polling and other heavy talking apps, and went from a soupy network full of squishy errors to now a fairly robust and reliable setup. That even included tinkering with my wifi setup a bit too. Lots of variables is all I am saying.
V2 is nothing but building blocks at this time, and I will give them the 3-4 months for them to show they can now iterate the software well, and quickly, now that this massive hurdle (which hardware is) has been eradicated.
This is what you had said about those who pre-ordered the new hub, however I ordered from the 1st promo deal (the customer application deal) at 10pm EST on August 17th and my shipping status is showing as âbackordered.â Did hubs actuly get shipped out as you said and if so how is it Iâm from the first promo deal, that you limited the amount of people who got it, and I am being told my order is backordered?