The lack of an authentication template kills projects before they start

[strich]

I'm an experienced C# developer who dabbles in ASPNET from time to time and I've created a number of complex projects in the past. On the whole I've enjoyed the experience and very much prefer to continue using C# for such projects.

I thought I've give Aspire a go and I want to love it, but the lack of a template project with authentication (Keycloak, etc) has almost killed my side project before it has begun. Between trying to use some Googled blog posts and the eShops example, I'm effectively lost. I get auth working, but various parts of it fail in different ways. I feel like I'm chasing my tail.

The root cause here is obviously my skill level - webdev isn't my day job - but even so the complexity is quite high in Aspire. Much more so than a traditional ASPNET monolith.

The eShops example is far too large with extra details and application features to be of use to someone who wants a template to build upon - I think I'd spend as much time unwinding it as building from scratch.

We really need a starter template with auth built in. I can imagine there are quite a lot of developers whose minimum bar for their project is an application with social login features, and right now I would guess a lot are bouncing off of Aspire due to complexity around it.

[davidfowl]

The complexity isn’t higher, it hasn’t changed at all. I’m not against an auth template, but did you use the exists.net core auth template?

Aspire doesn’t help you but it shouldn’t be any harder than it previously was.

PS: Auth is hard and want to do more to make it easier but we haven’t don’t that work yet.

[strich]

Sure I'm willing to accept it as true that auth isn't harder. But it is different through the change in tech stack from my own pov, which incurred a decent amount of frustration as my expectation going in to Aspire was that there would be a base level of features in a ready-to-go template. I'm happy to read guides and learn new concepts, but the goal was to kick start a little side project and this got in the way of that - the complexities of auth I hope we can agree aren't important to know at this stage of a side project.

The point, therefore, of this discussion is my assertion that the lack of template with these features hurts adoption of Aspire for the class of developer such as myself.

[davidfowl]

You can make that claim about lots of missing features and aspire based on what expectations a user has. The most upvoted issue in this repository isn't a missing auth template, it's missing support for Kubernetes (previously it was azure functions support). I'd encourage you to look at the issues sorted by upvotes and find the authentication one https://github.com/dotnet/aspire/issues?q=is%3Aissue%20state%3Aopen%20sort%3Areactions-%2B1-desc

From looking at user feedback IMO the top adoption blocker is EF migrations and how difficult they are to use in aspire.

Like I said, we want to make auth easier (for some definition of easier), but it hasn't bubbled to the top of the priority list yet (and it's not even clear what that means). Perhaps we can start with a sample in https://github.com/dotnet/aspire-samples/tree/main/samples

cc @DamianEdwards

[strich]

You would know a lot better than I about what your community is most concerned about. However I would like to point out that you may be at risk of only considering the funnel of users who are already part of the ecosystem and interested in improvements that concern them. Certainly you would not expect templates to be important to anyone who has gone far enough to care about "durable functions".

What about the funnel of users who come to Aspire for the first time and in day 1 bounce off it as they attempt to wrestle with the basic needs of their project that should aught to be available in a template? I can't give you any data to back this up, but my experience and my instinct is that the lack of templates is hurting uptake.

Anywho I think I've no more value to add here so I'll bow out. Thanks for talking David and looking forward to see Aspire progress!

[davidfowl]

This is the art of product making, not just feedback on the issues (if we were doing a burndown then this list wouldn't exist). The reason we even have a keycloak hosting integration was because we dabbled in this area. My main point is that auth is not worse than it already is. It's just as difficult and there are existing auth templates for asp.net core, aspire does not replace those (nor does it make it easier).

[oising]

@strich You're working/thinking at the wrong level of abstraction. Individual apps implement authentication. Aspire helps with distributed application building, which may involve sharing service names (i.e. dns) and/or ports between apps and services. If you don't know how to configure an individual application with an authentication mechanism, then Aspire isn't going to help you.