Run a business? Especially an online business? This is for you — you’ll get a commiserative kick out of it. Others may read it for the voyeurism — “Better him than me!” — or because you care (hi, mom).
Entrepreneurs running small businesses often indulge in black humour about how all-consuming a lifestyle it is. Some times it’s a blacker humour than other times. (And the cracking wise often stops altogether if the business dies — some failed ventures are really grim, and never suitable for look-back humour.)
My week-long Amsterdam holiday has been stained by server trouble. My website — PainScience.com, source of my income, product of a decade of obsession and passion — just keeps kacking out lately, unavailable to customers and readers, old and new, for hours at a time. This an utterly un-ignorable problem, and far worse still to try to deal with from the other side of the world, through nine time zones, and janky wifi access.
It’s not an unmitigated disaster. A few hours downtime is no business killer. But it certainly is a fun killer. It has been deeply exasperating. I can’t sell books without my website; I can’t pay rent without selling books; ergo, when the website is down I get extremely, er, let’s call it “concerned,” because the math is simple and brutal. Waiting for it to revive feels like waiting for the end of an episode of asthma. I’m embarrassed to admit just how emotionally intense it has been.
In a decade, I’ve never had site downtime exceeding a couple hours, or that needed more troubleshooting than simply shooting the hosting company an email and getting them to flip a switch, restart Apache, something trivial. But in the last month, I’ve had several — a half dozen? — downtime incidents of greater than 2 hours, and a couple over 4 hours. One was actually never solved, and I escaped it by actually evacuating my site to another machine entirely (which, by freakish coincidence, just happened to be ready and waiting because I was already planning on moving for other reasons).
The nadir of this ordeal came just hours before catching my flight to Europe.
I woke to find the site unresponsive yet again (less than a day after being warmly, articulately reassured by an engineer that all was repaired settled enough in time for my trip). That incident was resolved almost barely in time for me to be able to leave for the airport knowing that I still had a website and a functioning business.
Mere hours after arriving in Amsterdam — with less than 4 hours sleep in 48 hours, killing time walking around in the rain until hotel check-in — I found an unsecured wifi connection, checked in on the site, and to my amazement and horror … nothing! Again! Or still? Imagine yourself in this sodden, exhausted, far-from-home state. How much would your frustration would be compounded by a slow, unreliable internet access? Is the site really down? Or is it just a bit of twitchy Dutch wifi? At a moment like this, the clever isup.me service is invaluable: it confirmed “it’s not just you: PainScience.com is really down.”
That downtime incident lasted a good 3 hours — yet another improbably long repair.
And another odds-busting multi-hour failure came a couple days later. The engineer responsible, the same decent fellow who had reassured me that all would be well a few days earlier, wrote to me personally to explain what happened:
Purely human error. And entirely my error, so feel free to yell at me. It took so long to come back up because my blunder was substantial enough that I had to drive to the datacenter — my punishment for trying to make some progress in the middle of the night.
Server functionality and hosting services are usually pretty anonymous and nearly invisible: a black box from the user perspective. It all happens in a datacenter far, far away, lovingly (?) maintained by people you will never meet. It works 99%+ of the time, and rarely does the customer care much about the hiccoughs, or in many cases even know about them.
So after all I’ve been through lately, I was quite grateful to receive that surprisingly personal acknowledgement of engineering responsibility — and also a bit disturbed. The fate of my business in the hands of an earnest geek who made an honest mistake and had to drive to the datacenter in the middle of the night to fix it? Really? Wow. What if he was a little less earnest? What if he wasn’t paid well enough to care quite that much? What if he’d had a car accident on the way there?
But what did I think? Website hosting services are run by infallible magic elves?
Despite being relieved to know that my hosting service seems to be run by real people who care, I cannot even begin to trust that service now, not for a while, not until it’s gone a few weeks without any more failures. Many times now the improbable has happened, and the site has gone done harder and longer than any reasonable prediction. This intermittent negative reinforcement has forced me into a posture of intense, bleak vigilance, constantly foraging for unsecured wifi networks along the canals of Amsterdam — there’s damned few, by the way — checking the site as often as humanly possible, like a nervous tic — reload, reload, reload! — at a time when I should be enjoying Dutch beer, bitterballen, and special brownies.
This is certainly the most extreme example yet of how grinding the entrepreneurial lifestyle can be, the most challenging it has been so far.