If your website’s performance worldwide means anything to you, web hosting reliability is extremely crucial. I can’t emphasize this enough.
At the moment, I’m not concerned with monetizing my site (Ghost32writer.com)…but that doesn’t mean ultra-reliability is not important in my world. It most certainly is, and yes, I do feel that DreamHost (the web host for this site) provides that top drawer reliability just as WordPress provides the software power to produce pages worth publishing.
But there was a little glitch last night–just a little one–that showed me just how powerfully a site can be impacted by the slightest boo-boo.
Prior to writing here full time (beginning in October of 2013), I’d never had the kind of solid backup DreamHost provides. I did write for nearly six years at HubPages, and they weren’t bad in the reliability department, but there were times they’d be down for maintenance, pages didn’t load all that quickly (not bad, but not good, either), and for whatever reasons, steady viewership growth seemed to be out of the question.
Aside from HP, I tried half a dozen other web hosts over the years…with consistently dismal results. In the worst example, it took me 24 hours to figure out the host was lousy, 30 days to close my account, and even now (two years later), I still get occasional solicitations through both email and snail mail.
However, getting back to last night: Any web host can and will have a problem on occasion. It’s the nature of the Internet, combining huge numbers of users, super hi tech equipment, and of course hackers. There are so many moving parts to the Net that it’s not surprising when a glitch occurs; instead, it’s absolutely amazing that the thing works at all. So, when I lost connectivity to the site (G32) around 8:00 p.m., I didn’t get too excited or panicked about it…but after it went on for more than twenty minutes or so, I did begin tracking down the problem. If I couldn’t access my site, that was one thing. If my readers couldn’t find it, that was another matter entirely.
As it turns out, they couldn’t, but I didn’t know that for sure.
As a way to check, though, I did email my #1 reader, Becky, asking her if she could access the site. I didn’t hear back until after the problem was resolved, but it was something to do.
In the meantime, the following check-steps were taken:
1. I tried accessing the rest of the Internet. Yep, no problem there.
2. Next, could my wife’s computer get to the site? No, it could not.
3. Going to the Command Prompt, a ping was sent. Couldn’t ping it. Aha; now we knew the problem was “out there” somewhere.
4. Back to Command Prompt, tried a traceroute. Nope; at four or five jumps after leaving the Cox network, the signal died. Hmmm….
5. Between one thing and another, and stopping to ponder (i.e. playing a computer game of Hearts here and there to kill time), a couple of hours had passed…and it was time to email DreamHost.
The following series of messages ensued:
Me to DreamHost:
> I’m a satisfied customer, loved DreamCon (meeting Simon, Dallas, and the
> > rest), but now have a (minor?) problem. Don’t know that you can help, but I
> > lost connectivity to my site (Ghost32writer.com) a couple of hours ago.
> > Traceroute shows my attempts to get there do get out of the Cox network (my
> > provider) by several jumps. My provider tells me it’s breaking down at an
> > unnamed router at 184.108.40.206 but neither he nor I have a clue as to whom
> > (or where) that might be.
> > He also tells me that indicates a “probably fairly major break somewhere”,
> > which would make it likely I’m not the only one with the problem, and that
> > somebody is already on it–but we don’t KNOW that for sure.
> > Any insights, suggestions, and/or magical waving of wands would be utterly
> > welcome. The site’s Alexa rankings have been rocketing up the charts
> > during the three months I’ve been writing there full time–from 23.8
> > millionth in early October to 3.99 millionth yesterday. I’d hate to lose the
> > momentum because I was slow to ask for any possible help.
> > Thanks,
> > Fred
Me to DreamHost:
> > Ghost32writer.com)…and seconds later, connectivity is restored!
> > Magic. It must be magic.
> > Double, maybe even Triple Thanks,:)
> > Fred
Me to DreamHost:
> > Connectivity gone again! (Which I’m guessing means somebody is at least
> > TRYING to fix the problem–LOL!)
> > Fred
All of these three messages took place between 10:00 and 10:30 p.m. By 11:00 p.m., connectivity was once again fully restored and stayed that way. Then, at 3:46 a.m., in came the following message from Daniel H. at DreamHost:
DreamHost to Me:
> My apologies for the trouble.
> There was an issue in our Irvine data center which affected your site.
> It has been resolved, so everything should be good now.
> More info here:
> Thanks for your patience!
> Daniel H.
> DreamHost Support
To which I responded:
Thanks. You guys are freaking awesome.
Okay, so why do I say ultra-reliability in one’s web host is ultra-important?
It’s like this: I watch my site’s stats like a hawk about to swoop down on a mouse, not necessarily during the day, but always during my nighttime writing hours. Prior to the glitch, which lasted for roughly three hours before it was totally fixed, my “human visitor” count was on track to hit around 2,200 for the day–but with three out of the day’s last four hours eliminated, the actual tally ended up being just over 1,800. It’s recovered nicely today, and I couldn’t say the lost connectivity time cost me much (if anything) in readership momentum.
But it surely did cost me at Alexa.com. Prior to that three hour gap, the site’s Alexa rankings had been climbing steadily for the last three months, up from 23.8 millionth in the world to 3.99 millionth in just ninety days–with never a single drop in the rankings during all that time.
Makes a webmaster feel all bulletproof and stuff, huh?
I’m kidding; there might be something bulletproof out there, but it’s not the Internet–unless you’re Google, maybe.
Today, the site’s rankings took a real hit, dropping to 4.5 millionth in the world just that fast. Half a million places down in a single three hour period.
Alexa claims it looks over your site for the “past three months” to decide where you should rank, but there has to be more to it than that. At a guess, their bots tell their computers, “Hey, this site was inaccessible for three full hours. Buddy, that’s like an eternity worth of milliseconds–penalize it! Now!”
Or maybe you, being a super savvy reader, have a different theory on how that works–but I do know one thing. This sort of glitch is incredibly rare with DreamHost, and yes, they were on it like a duck on a June bug…but what if I were still trying to make a go of a site with a less reliable host?
That would be bad. That would be very, very bad.