[This is a suggestion for Wyze back end operations. No other forum category fit.]
One compounding issue from the last few outages (Amazon AWS and otherwise) was all the pent up notifications and rules and triggers that got queued and had to be executed late when services were restored. The question is: why bother?
Probably obvious and already under discussion but wouldn’t it be better for everyone to simply discard those queues? People seem to still be getting some delayed notifications / announcements a day later. They are worse than useless; they are misleading.
Whatever someone wanted to happen - whether opening a live view or unlocking a door or turning on a light or getting notified of a vehicle - becomes mostly irrelevant 12 hours later.
I wonder if Wyze can create redundancy instead so that there isn’t a backlog in the future. For example, spamoni4 wondered if they could have an Eastcoast server for those on the eastcoast and a westcoast server for those on the westcoast, then if one goes down it automatically falls back to the other one with little to no interruption. Then we shouldn’t have that same issue.
Otherwise, perhaps there can be a set interval limit. If there is an outage or delay lasting longer than X amount of time, then dump/clear everything before that time. It seems reasonable to leave a backlog within a couple of hours or so, but not after 12 hours.
I believe most competitors just cancel all notifications that happen during any outage, and thus their downtime is significantly reduced, at least based on the outage report graphs I viewed the other day. Several other camera companies had outages at the same time as Wyze, but they recovered more quickly, I think this is mostly because they just dropped all activity during the downtime. There are plusses and minuses with this, but I think something in the middle would be good…clear anything that is X amount of time old, but allow newer activity to run as normal. Something like that seems optimal.
What’s the breakdown, 60% plus, 40% minus, or…?
From a completely ignorant POV (mine) the current ‘plan’ just seems abundantly dumb.
I don’t know that I would quantify the plus/minus ratio as much of it depends on each individual’s situation. Some notifications would be really valuable later (ie: tell me that my HMS alarm triggered or that a cam saw an unknown person on my property, or that there is/was a leak on my leak sensor…even a few hours late that notification could make a huge difference compared to never being notified. There are countless automations or notifications people may like to run, even if they are late.
On the other hand, many notifications are fairly frivolous. Sometimes I have cams tell me when they detect my cats in certain areas for various reasons so I can see a thumbnail/screenshot and easily swipe it away without even viewing the whole event, or when certain doors open. I don’t need to know about those things hours later.
So on the one hand, getting some critical notifications even late can be really important, while on the other hand, most of my notifications are frivolous. But others might have ONLY critical notifications, so need them no matter how delayed.
Thus hard to quantify and generalize for everyone. Seeing as how almost all competitors appear to simply empty their entire notification backlog, there is a case to argue that Wyze doing something different could be a good business strategy. On the other hand, there’s also probably a good reason why almost everyone else just clears their queue and doesn’t pass on any backlogged notifications, because the alternative is problematic.
Just saying there are positives and negatives both ways, but I think the general market prefers as much uptime as possible, and notifications come second to that. Wyze can always send out an alert that there was an outage, and this prompt should let people know to review the status of their devices: make sure there aren’t any active leaks on leak sensors, load up the events screen and sort by person detections to see if there were any concerning videos detected recently, check whatever else could have been affected with notifications or rule automation. Having functionality back up earlier with a notice that there was an outage to be aware of would be preferable to most people.
I say dump notifications and shut down all future functions until after the outage. You know there were millions of people trying to troubleshoot their devices that didn’t know about the outage, cycling power on and off, deleting and reinstalling the app, removing and installing devices, etc. How many hours did those people add to the outage?
I would never…I value your job security in that department too much my friend.
(for the casual observer, peep and I are buddies, and I"m just joking with him)
Don’t worry about charting stuff, some of us analytical masochists actually enjoy doing that kind of stuff, so you can leave that dreadful spreadsheet stuff to us.
And now I’m going to go set up my 5 new Wyze Lamp Sockets that just came in! YAY! If I can use them as I expect, I’ll buy a bunch more to automate a bunch of lighting in my house and have to update my spreadsheet on how much Wyze stuff I have so I can keep beating spamoni4 & R.Good for having the Wyze-iest smart home…so back on topic…having a back log of a few notifications a few hours old isn’t a big deal, but when you have 150-200 Wyze devices with dozens to hundreds of notifications hours or days old…then this wishlist item starts to matter a lot more.