This has been discussed many times in the forums. I don’t want to repeat myself too much, so I will say that the first thing is to be honest with yourself about what your own priorities are. Every system has pluses and minuses.
SmartThings’ biggest minus is instability – – they keep changing things, and when they change them, sometimes they break. There may be workarounds, there may be simple fixes, the glitches may be minor – –but they do keep happening. And it’s not because of your configuration or the strength of your mesh or anything else within a customer’s control. Stuff which works on Monday will stop working on Tuesday, because the SmartThings’ design philosophy continues to be “run fast and break things.” They’re always sincerely sorry and they always try to get them fixed again as soon as possible, but it’s just the way they operate.
The flipside of that is that it’s one of the most exciting systems in the very low cost range, it’s always adding new features and new capabilities, and at least at the present it’s very open to allowing customers to add new features as well. WebCore is the perfect example. It is an amazing project and adds a ton of functionality.
But if the time has come when you want to trade that kind of versatility for a more stable system, there are certainly alternatives. There aren’t any low cost alternatives that have the same variety of devices and rules options for nonprogrammers that SmartThings has. But if you just want a system where the lights come on at sunset every night and a smart door lock and a few motion sensors, there are definitely options.
If you’re OK with an extremely limited set of Devices to choose from and single level if this then that rules, I personally like and use HomeKit. When used with Lutron switches, the Philips Hue bridge including sensors, and a couple of other brands, it’s easy, reliable, low maintenance, and essentially a “set and forget” system.
If you want to be able to use your existing Z wave and zigbee Devices, there are still some options, but again, simple rules, and many may limit you to only 40 total devices or so.
The following thread is old but most of it still applies:
Samsung is planning to move us to an entirely new architecture sometime this year. While the end result may be great, it’s hard to imagine that the transition period will provide more stability than the current system.
Different systems are going to work for different people. Your own household’s tolerance for unpredictability Will depend on many different things. How many people are in the household? How often are they home? How burdensome is it if the home automation doesn’t work on any given day? What use cases are you trying to accomplish and why? How much money are you willing to spend? How important is it to you to protect the investment you have already made?
There’s no one right answer. There are a dozen different right answers, that fit a dozen different preference profiles. 
So when do you throw in the towel? That answer is going to be different for different people as well.
You may also find that you only want to move some use cases to a different system and keep SmartThings for use cases where reliability is less important. If you already have Lutron Caseta switches and the Phillips hue bridge, and you already use an iPhone, you could move all of your time-based and many of your sensor-based basic lighting scenarios over to HomeKit without adding any new equipment at all, and greatly improve the reliability of those use cases. And then just ride out the transition and see if the new platform is indeed any more reliable.
Not an easy question, and there are no easy answers. But there are enough choices out there now that you should be able to find some answer that will work for your household as long as you can be honest about your own preferences.
My own smart house is pretty dumb right now, which is not what I would’ve thought I wanted. But it works reliably, it meets my needs, and I stay within my budget. Choice is good. 