The Yonic Corner is now self-hosted
Previously, The Yonic Corner was being hosted on Netlify —and Github Pages in the case of the Legacy version. It used to be pretty convenient to write content and deploy it there, but as I was approaching certain limitations, I decided to make the switch back to a self-hosted environment.
The performance of the blog might not be on par with when it was on the cloud, but I’ll be working on optimizing it little by little.
Running out of space
Lately I’ve just been worried about running out of space altogether on any digital storage I own, honestly…
The main problem that I had was that GitHub Pages has a hard limit of 1 GB in storage. For most small scoped sites, this shouldn’t be much of a problem. But in my case, it was my personal blog, and while text is pretty easy to compress and make thousands of pages without even coming close to that limit, media could potentially do that.
Especially if we’re dealing with optimized responsive images! Most images in the modern version of my site come in multiple sizes for PC and mobile screens, and multiple formats for backwards compatibility. Add that to the fact that Astro keeps a copy of the original image, and it’s a recipe for a bloated image cache, some of which wouldn’t be used at all.
An image CDN could work marvels for this situation, but I was a bit skeptical on those to work well, especially because of the reason below.
Watch out for HTTPS!
In a previous post, I talked about how pages have begun breaking because HTTPS support for legacy systems is being dropped. I’m expecting that GitHub Pages may be doing the same thing in the near future should they begin enforcing mandatory HTTPS on all their pages. This would render the Legacy Version of this blog useless. It was designed with old operating systems in mind, after all!
And this also applies to what I was talking about before: Normally I would be okay with just using an image CDN, but all of these networks work under HTTPS in a way that only modern systems can understand.
This has eventually forced my hand: I needed to self-host my own blog. That way, I’d have full control over what I want to do. I could even set up a cipher suite specifically for the legacy blog, so you could browse the legacy blog in HTTPS while in an old system!
But is HTTPS really that necessary here? This blog is nothing but a static site. You’re not really sending anything to the server other than plain GET requests, and the contents of this site aren’t the slightest sensitive nor illegal.
If privacy is of concern, you could just set up a Tor network connection and be done with it. But again: Static pages can’t even use cookies nor session tokens, at least not directly; if the request doesn’t come with anything regarding authentication you won’t be able to learn that much about your readers anyway.
The blog also works as a reverse proxy server to another program that optimizes and caches images on demand, so I can show the images regardless of whether you’re using regular HTTP or HTTPS. Plus, I can shove in a storage bucket to have virtually limitless —and cheap— storage space for my blog.
However, this is not without its drawbacks. Since I’m hosting it myself, all maintenance tasks are on me as well. I still have work to be done to optimize the server, to automatize updating the blog, and to tidy up the API for remote images, which is a total mess right now, let alone a rudimentary solution. So until then, do expect some slowdowns!
