Sentry just gave $750k to open source projects(blog.sentry.io) |
Sentry just gave $750k to open source projects(blog.sentry.io) |
Sentry is a multi-billion company, and Open Source pays exactly none of their bills (though it may serve other purposes). This leads to pages such as [1] where they actively steer users away from self-hosting their "open source" (in name only) solution.
Much has been said about Sentry's switch to "Fair Source"[2], but for me personally, the ship of "open source in name only" sailed long before that with the ever-increasing complexity of managing your own setup.[3] It’s clear that the priority here has shifted to pushing users toward their hosted, paid plans. Business models beat licenses every time if you want to understand actual intentions.
Disclaimer: I am the solo-everything at a competitor, which is in fact _not_ Open Source.[4]
[1] https://sentry.io/resources/self-hosted-vs-cloud/ (click on pdf for scary pictures)
[2] https://news.ycombinator.com/item?id=41171665
[3] https://github.com/getsentry/sentry/ commit/78bc759d1be4fa6b8ae3e2764e7156e05eb22ab9
The point about "open source in name only" due to hosting complexity overlooks that this is a natural consequence of the product's evolution, not malicious intent. Enterprise-grade scaling of such monitoring requires sophisticated infrastructure. Suggesting they should artificially keep it simple for self-hosters would hold back product development.
I agree about the real, tangible benefits for open source. But wouldn't you agree there's at least _some_ point on the scale of evil where donor motivations can be questioned?
It's definitely marketing, otherwise they wouldn't be publishing the blog article - but what you're insinuating doesnt hold up considering the timeline
They must have the most inroads there, I imagine.
That is also why Rust and rrweb are overrepresented.
But to answer your question more realistically:
* Familiarity
* Easy/built-in admin
* Longevity
I would also never consider switching an existing app from one to the other.
FastAPI make sense only if you write an api for a third party frontend or service.
But what is the sense to compare on a single feature? Obviously, Django doesn't cover 100% of use cases but, usually, you choose a technology that simplify you live overall not on specific aspects.