Site Insight: The Downtime — What Happened

So, perhaps not all of you, but a lot of you saw that the site was up and down, but mostly down on Friday and Sunday. This post is to explain what happened.

Now, there’s two ways I could go with this post. The first, by laying out in exhaustive detail every step of the journey the site and I went on. Or secondly, a brief summation — though fans of my work know that I am hard-pressed to accomplish brief.

For your sanity and mine I will choose the latter.

Basically, fights.sexy is a very large site — 326 pages as of right now. A large site that is feature-rich. With a forum, a chat program running on every page, the sidebar that contains a bunch of items like a display of our tweets, links to both random and our newest stories, recent comments and forum posts, etc….

Each of those elements makes the site more difficult for a server to run, ESPECIALLY on WordPress.

Why? Well, apparently (I know this only from what I am told by techs who I am trusting to tell me the truth), WordPress (which the site is built on) is a “make-it-work-first” platform. Meaning, that plugins and site function are designed to work, and then optimization of them is forgotten. That being the case, WordPress sites are super prone to just gobbling up server resources, even if they otherwise shouldn’t or wouldn’t, if they had been correctly optimized.

So, what does that mean? It means that as the site grows, the server it runs on is under more and more strain. Not because the user is a burden, but because WordPress’ unique design is. With each user posing the threat of their instance of the site sparking into an inferno of un-optimized code.

That threat I worked to contain as best I could a month ago, by removing less-than-useful add-ons, adding caches, and stripping the images on the site of their meta-data. Along with that, I also upgraded our site’s server when we first had the OVERLOAD issue.

Despite that, on Friday, it came again. Why? I really don’t know. I had asked the techs at my host if it could be the near constant stream of Russian accounts being created on our forum. Their response was, basically every site faces the onslaught of Russian bots. And that generally, they only increase load by about 10%. Where as, on Friday, our site was 200% of the capacity allotted to it. Meaning, the bots, though annoying and malicious, weren’t the root cause.

Regardless of cause, however, I had to fix the issue. And after basically half-a-day of silence from my host, turned out to be paying them for a top-end dedicated server.

Expensive though it was, I made the payment to get the site back up. HOWEVER, when the techs tried to transfer to the site to that dedicated server on Friday night, the site broke, as all the data tables were corrupted — an issue that surfaced to users when they could not comment.

Why were they corrupted? Well, it seems that WordPress, once again, placed a poison pill in the code (something they did before — thereby forcing me to have to rebuild the entire site when we moved from WordPress’ hosting services to this new one’s). The poison pill was not something my new host could fix and so … basically, they and we had to abandon the new server and remain on our old server.

Now, our old server was not weak or deficient in anyway other than the fact that we were sharing it with others. Others which the host moved to make room for us to exist there on our own. Giving us the full resources of the server and not just an overload-protected slice of it.

That moving happened Saturday, about mid-day and since then the site has been working perfectly. The host assured me we would no longer have overload issues, as we had in the past and specifically on Friday.

And so, to make a long story long, at least for now, we’re back on the road.

With all of that said, I want to say thank you.

To all the people who DM’d, emailed, and elsewhere messaged me. For the support those people gave to me as I struggled to get the site back up. And to all of the rivals who are so into the site that when it goes down for even a moment, worry along with me.

Your words, encouragement, and engagement mean more to me than you can know. Thank you so, so much.

1 thought on “Site Insight: The Downtime — What Happened

Leave a Reply

Your email address will not be published. Required fields are marked *