Is .NET Core stable enough for production?

Yes, you can say yes. But with some care.

First we would have to define what is to be "stable for production."

When Microsoft released version 1.0 of it, it rated it stable for production . So objectively and officially we can say yes, it's ready to be used in commercial applications.

In fact a lot of people are already using these conditions.

Some people complain about some things. Including the lack of some components. Some that will never be implemented because it does not make sense in .NET Core (but even that has changed a bit), others that need to wait a bit, or get some alternative, as many people have done.

For example, the Entity Framework Core will be much better when complete, but now there are several things missing. Some will not because it has never been appropriate in earlier versions, such as model first .

It works differently than the .NET Framework, so for those who are used to it, it can be a bit difficult to adapt. Some things may not work because the person does not know the new way. Other things may not be working well at all. But nothing critical that you can not use. Of course, it might give a little more pioneering work ( early adopter ). But there are no bugs compromising absurdities. (Well, that's kind of overdone, it was worth more when I originally wrote.)

There may be a little tool missing. There are already many famous libraries ported to it, but not everything. But less and less.

We already have updates that have improved some things. When 2.0 left , much more stable yet much more complete was closer to the level than everyone wants. Visual Studio 2017 already comes with much improved support for it, and there will be updates over time. The new one is already on the way to passing the support that the traditional framework has.

If you're bringing something old to .NET Core, I do not know if it's the best idea right now. And depending on what you use, or whether it will be interesting at some point. Something needs so much adaptation that it only pays off if it does from scratch, which many will say is crazy (actually it depends on the quality of the old, the size, and other less technical issues).

At the bottom only you can respond if it is good for what you need. Have to start wearing and seeing. There are people with experience who tried to use it, found some obstacle and gave up for an hour, there are those who have managed to get along without even understanding much of it. So the experience of each is unique, and depends on the commitment put and need that the person has.

Many people find PHP 7 stable. For me it does not roll, I would have more difficulty porting my PHP applications to 7 than doing a new application for .NET Core. But I can not say that people who find him stable are crazy, ignorant, or something, depends on each one.

Recently I was in the Microsoft MVP Communuty Connection and had a lot of MVP already using in production successfully.

When 2.1 came out practically completed what was expected of 2.0 and is already quite complete in its core (no pun intended) and what was missing comes in the next 2.2 . Of course ASP.NET still lacks a bit more, but little, EF is a bit late and has a different intent, but it already has a lot of interesting new stuff that classic EF does not have.

In 2019 comes the 3.0 that I think completes the cycle and even meets almost all the needs that the .NET Framework met, including those specific to Windows through the compatibility kits (it will not be part of the basic distribution, but it's super easy to insert in the application). What we do not enter is because it is not to use even. And look even Windows Forms will be supported. You can even use EF 6 for those who need something unique to it that they can not give up.

My vision, with no insider information being Microsoft MVP, is that the .NET Framework will become secondary and will receive less update by political decision, but also technical, which I think is beside the point in this question. Not only will every investment be in Core, but the community is working very hard on it, there is no way to do this in the Framework for reasons that are not relevant here.

So I tried to respond objectively and not try to fall into opinions, because each one has his own.