i wonder if you could make a small business out of providing Forgejo Actions compute resources for projects that want to pay for what they use instead of giving control up to GitHub
you should be able to have reasonably high margins because: (a) your customers are people who by definition are willing to spend resources on "not being on GitHub" (b) it's a small space that's annoying enough to work with (broken tooling, little volume) that even if you publish everything as OSS you're not going to get competitors undercutting you the moment you get any traction
the downsides is that it's a small space, the tooling is very broken, and expectations are higher for a paid service than a free volunteer-run one (though... frankly not that much higher, given it's OSS we're talking about here). also you'll have to use kubernetes
i doubt you could turn this into even one full-time position but if it's something you're already doing for other reasons (such as "you need CI") maybe it's worth it?
@whitequark GitHub CI really is the classic "undercharge for $thing until you've created a novel middleman in an existing market, then overcharge for $thing" strategy, but I think that gets lost because for public projects, "overcharge" comes more in the form of AI scraping and other bullshit, and less in the form of monetary cost.
@whitequark Would lock-in be lower with Forejo Actions, because if the customer gets tired of dealing with you they can self-host Forejo but not Github? That might be an incentive to some customers.
you _would_ be effectively competing with every volunteer-run CI service out there, but they're so understaffed that it's not hard to do better when you pay attention to the project, and also it could be seen as mutually beneficial: you get customers out of it, those services get to redirect people who want a resemblance of an SLA away and not have to deal with them
@xgranade i knew this in my bones back in... ugh i don't even remember when, 2022 at least? i had vague mental plans of migrating Amaranth away that i had to put on pause because of how costly it is to dothat
probably the #1 risk from a business perspective would be one of the big services like Namespace, Blacksmith, etc adding a Forgejo offering. it would likely be so much better technically you'd be left relying on customer loyalty. probably doesn't make sense to develop anything new if/when that happens
@whitequark I think you could get a lot of customer loyalty from providing a hosted service comprised of open source components. You could even have a guide for "here's how you could self-host" as an anti-lock-in guarantee. Showing the specifics of what you're paying for is also pretty compelling.
@jornfranke yes, for sure it would have to be a European datacenter (I'm not sure if you even want a cloud provider here; you can get donated bare metal for quite cheap and there are plenty of good colo opportunities)
@whitequark Many CI/CD offerings do not solve secure sldc and isolate their tenants (customers) properly (especially Github does not). Then sovereignty might be an issue.
You may also want to partner with European Cloud providers that are stronger on the Enterprise market.
@jornfranke I think you are definitely onto something here, and I'd expect B2B to be more pleasant than B2C, but also I personally for example have zero experience in that particular sector and I'm not sure that's a viable starting point
@whitequark Yes, but it cannot harm to partner in addition with a cloud provider targeting enterprise sector. If justified or not - the enterprise sector has some specific requirements (e.g. reliability (multiple DCs), encryption, sovereignty, strong isolation of tenants etc.). They are very specific and not fully easily solveable oneself. It offers so additional premiums.
@whitequark@mastodon.social I hope so, because this (and it then hopefully partially funding some generic public runners too) is on my mind as a project for next year.