This post is part of an ongoing series on the best practices for effective and insightful database monitoring. Much of what's covered in these posts is unintuitive, yet vital to understand. Previous posts have covered Why Percentiles Don't Work the Way You Think; how to tell If a Query Is Bad; an explanation of why, when looking at charts, you should understand that A Trendline is a Model; and Why You Should Almost Never Alert on Thresholds.
Imagine you're an up-and-coming retail site with ambitions to someday become the world's next Etsy or Amazon. So far you've had promising early success, but you're still growing; your userbase is relatively small, but admirably devoted. After many months' long hours of meticulous coding and visionary marketing meetings, you're comfortable with the core of your enterprise and confident in both its concept and execution. In your mind, it's only a matter of time before your site blows up -- you know it offers the best retail experience available, with a cleaner, more pleasing interface than competing retailers, plus better selection, the best site performance for your current level of traffic, and stellar customer service.
One grey, winter morning, with no warning, a product that's only available on your site is featured in articles on a slew of buzzy, twittering news outlets and blogs; suddenly, millions of Facebook posts appear that link to an item that can only be found on your storefront. It's gone viral, truly viral. With this opportunity and attention, it's suddenly your website's chance to make its mark.
Within minutes, the new traffic and queries overload your systems; an issue manifests in the server. Up until this point you've never needed a monitoring tool, so you've never bothered to get one. But now there's an issue in your system, and even though you can feel its harmful impact, you can't see it -- it's impossible to pinpoint the source of the problem, and you're left to do nothing but watch a train wreck in slow motion: your site goes down for the rest of the day. Potential customers find the website broken and they leave, frustrated; with fickle shopping habits and low tolerance for broken webpages, they'll probably never return. Worse, the product's manufacturer is furious that your site can't keep up with demand. They immediately jump ship and find a different retailer -- one of your competitors -- who can handle the traffic. All the blogs that originally posted about the hot item change their links' URLs.
Opportunity severely missed. Reputation tarnished.
This story illustrates just one reason why it's important to think about database monitoring even if you've never had a database-related crisis in the past. (It's a realistic scenario too: the Slashdot Effect, a.k.a. the Reddit Hug of Death, is a common phenomenon, occuring when "a popular website links to a smaller site, causing a massive increase in traffic. This overloads the smaller site, causing it to slow down or even temporarily become unavailable.") But issues can arise even without such scenarios -- databases are inherently complex and difficult to watch without powerful assistance. When does an organization cross a threshhold and suddenly find itself in need of monitoring capabilities? It's a question that needs to be considered. The truth is that if your organization is growing, it's not a matter of if you'll ever need monitoring, it's a matter of when.
This is not fear-mongering. In fact, it's the exact opposite: establishing database monitoring practices with proven tools shows clear-thinking and prudence, not panic. In reality, all organizations who rely on their databases will face a moment when it will be necessary to increase the level of insight into their systems. Imagine a doctor telling a patient, "Now that you're of a certain age, you've got to start paying attention to your heart health before you have a heart attack, not after. Not because you've ever had cardiac issues in the past, but in order to avoid them in the future." It isn't just a recommendation, it's a smart and healthy practice that will have real, beneficial effects.
So, when is it too late to monitor? Well, as the story above shows, it's almost impossible to predict the exact time or scenario when you'll first need it (although for retailers, Black Friday is a perennial case), but if you're an organization that uses database systems and you have ambitions to grow, the moment for monitoring is guaranteed to come. And there's no reason not to get started (and also begin reaping the many benefits beyond crisis-aversion) -- it's fast and easy to set up a free VividCortex trial. It's important to start monitoring now, so that when the moment arrives, you'll have the success story you were meant to -- not a lost opportunity, or worse.