Is Addressables Production Ready?

The title says it all. Please let me know what you think.

 

Could you describe a little about your use-case?

There's some stuff in the package that work well, some that work less well, you could open up the source code for the package and get a pretty good grasp of what stuff is really janky.

In my personal experience, as long as you don't have some really odd stuff that you want to do you should be fine.

Edit:
Perhaps the best way to get a good clue of the state of the addressables package is the samples repository, available here:
https://github.com/Unity-Technologies/Addressables-Sample

 

Different projects - different experience

We use Addressables about a year for now and really tired. All wee need is load bundle and load asset from it thats it.
All bundles took about 650 Mb in our project and mostly every build they fail to build or fail to do something else.

Last LTS Unity just crash on loading bundles in our project. So we plan to go as far as possible from Addressables for now.

What will be your experience? Who knows

 

Oh yeah, should've mentioned that but yes I've also noted that about every second build fails.

 

Is this a joke? Obviously this means its not production ready then?

 

According to the manual page:
"Package version 1.8.5 is verified to work with 2019.4.", with verified meaning "Verified packages are packages that have undergone testing with a specific version of Unity, and all other packages verified for that version.".

 

Only If you have time and money to throw away

 

From my experience, just about. We don't use AssetReferences (I made our own version of AssetReference instead) so can't comment on that part (I think there have been a lot of issues with them in 1.15.1?) but for us it feels like we've retro-fitted (and actually shipped) a largeish project using addressables at the earliest time it could have been done. Like with a lot of Unity products, it is still very much in development and its development is being driven by real world use, rather than proper in-house testing. We've finally gotten to a point where we don't have to copy the addressables package to fix issues ourselves and have now placed our trust in unity that this is infact the future of asset bundle management. I do have a little anxiety that it will be scrapped and replaced with something new, as so often is the case with Unity.

 

I encourage everyone reading this post to go and give this post a +1. https://forum.unity.com/threads/please-open-source-addressables.970647/ If we can make pull requests, we can really push addressables to be a lot more stable than it is now.

 

I agree with everything written so far. It's almost there -- but we've lost countless hours to issues, quirks, lack of documentation, etc. I can't echo the thoughts from @c0nd3v enough; this is "permission-to-play" level stuff, and the fact that there are known bugs that didn't make the most recent release [at least some of which are >1year since first being reported, 1 of the reports even contains the code to fix the issue] is... flabbergasting, at best.

The only reason I'd use it for a production project today is if you're having slow startup times and need to migrate from Resources stuff; if you have a good AssetBundle workflow already, stick with that. Otherwise, contemplate it & maybe consider moving to Addressables.

I wish this wasn't true; this is core functionality and, for a company on the brink of IPO, the fact that some of this stuff hasn't been fixed or improved in months+++ is...

Well let's just say that, unless I was in charge/empowered to get my team to fix the issues, then I'd be selling as soon as my lockup was over! :-|