So, quick observation. I'm never going back to WordPress.

Yes, it works. But it's haggard with age and frankly hazardous to use. I'm not sure if you've read my previous articles, but I was getting an average of two hacking attempts every day from the People's Republic of China. Sometimes I'll wonder what the hell I did to offend the Red Army so.

Beyond that, Ghost is just a breeze to set up and manage. Sure, there's a few stumbling blocks. But it's rarely anything that will destroy your work at large. Since they're so stupidly easy to build with DigitalOcean or AWS or GCP, you can build them up & tear them down in record time. By the third go around, you'll know the drill very well.

This may be a quality that haunts all IT folk, but when I see wires dangling out of a panel or stretched across multiple racks it invokes absolute dread. The same goes with poor operations practice.

I mentioned the security details as well, which are awesome. This server is brand new as I write this article, and that was enabled due to Ghost's simple backup & recovery methods. One text file is spat out, the same one that goes into the fresh site. No nightmare plugins or eldritch backup solutions to haunt your dreams. It's such a quick & easy fix that I'm considering throwing out the key and recreating the entire site every time I want to make a change or publish a new article.

Wonderfully, these tenets stick well with the principles that I've tried to recommend in my work. Stuff should be designed with a lifecycle in mind: no machine should be immutable or considered the cornerstone of a business. These things introduce uncertanties in business continuation in the best of times and can even turn into security holes that gather undesired attention.

Thankfully this is more of a passion project than a core business, but it's worth doing things with best efforts at all times. This may be a quality that haunts all IT folk, but when I see wires dangling out of a panel or stretched across multiple racks it invokes absolute dread. The same goes with poor operations practice. So, yeah, don't ever rely on one big machine that might go down. Because one day, it will.

A couple of years ago, there was a hot new devOps philosophy in town. It was to treat your servers like cattle rather than pets. Since I was not yet a server monkey, I had no idea what this meant. But as the pile of container nodes on my desk grows, I am beginning to have an idea of what this means.

These things are born, they live, and they die. The best way to record what they are and what they do is with a binary data stream, like DNA. The best place to keep this is inside a repository that you have editing access to. This way, when they die (and they will,) you can have fresh clones ready and waiting.

Thankfully Digital Ocean and Ghost do lots of that work for me. I think the next project will be finding a way to build up this entire site with a single command. That would be metal indeed.

Before I sign off, I'd like to impart some love in my readers. You guys keep coming back and checking this site out and that really impresses me. Get in touch if there's anything I can help out with. And lots of love.