Why Archive Every Business Transaction?
The other day I found myself in a discussion with a prospect regarding the merits of capturing complete stack traces for every transaction that flowed through their production system. You see AppDynamics doesn’t do that; we capture problematic business transactions, and this bloke seemed concerned someone would one day ask him for the details of a specific transaction that our method would leave him underprepared to fulfill. As the conversation played on, we explored the cost – both literal, in the sense of hardware required for data warehousing, and virtual, like the compounding impact additional polling during bad times has on performance. Ultimately we arrived at the conclusion that understanding your haystack is better than having the haystack. This point can be subtle, so let’s explore how we got there…
Since AppDynamics automatically baselines business transactions, we can very quickly tell the difference between a normal transaction and an outlier, or in the context of this analogy: hay vs. needles. When we encounter a piece of hay we record the metadata of the transaction. You know: how long the strand is, where it resides, who put it in the stack, etc. In fact, the part of me that humbly struggles with homonyms likes to think of this as a piece of “hey.” I imagine someone (a business transaction) flying past me on the highway and waiving “heeeyy!” as I feverishly jot down some key details about their ride. For the most part I already know enough about them, they’re speeding along unabated, just like the plethora of friends to follow, no need to dig too deep, just note the event so we can better understand our “heystack.”
In contrast, when a particular transaction performs several standard deviations outside of the norm, or perhaps worse than a given threshold, we need to know more than “it’s slow.” This is where the context before content idea really starts to come into light. Because we know what a good piece of hay looks like, and because we understand how our virtual haystack looks at any given moment, we have the context to both appreciate and spot the needles – the need for content. Needles are special. They’re shiny. And you probably want to trade them for hay. So we need to know all their specs. We need content.
No worries, that’s why AppDynamics captures complete transaction flows for these items. But what if – returning to the highway analogy for a moment – a funeral procession (a bout of bad performance) drives by? The act of gathering information during this kind of needle storm could exacerbate the issue and result in exaggerated wait times, inaccurate information, and a unnecessarily pitiful user experience. Well we have the context to understand when that’s happening so AppDynamics will automatically throttle back needle gathering to ensure our application never negatively impacts yours. Remember the ultimate goal here is to ensure you have all the tools you need to deliver an exceptional user experience, in the short and long term.
Thinking about the long term for a moment … we should probably use this information to improve the performance of the application. So let’s pass the data about the problems along, shall we? But … if all you ever gave your team were needles, they might forget what hay looks like. This lack of context could hinder their ability to remediate issues. To mitigate these risks, AppDynamics will snag the occasional piece of hay… just for comparison’s sake and ensure your team, your developers, have all the context and content they need to improve the performance of your applications.
So as you can see, you don’t need to house an entire haystack to effectively identify and remediate problems. With a clear understanding of what the haystack would look like – a holographic haystack, if you will – a few pieces of hay for reference, and a spotlight on all your needles, you’ll have everything you need to see, act, and know more quickly than ever before.
Also check out how Citrix uses AppDynamics to capture business transactions, increase visibility, and lower their MTTR for their complex applications.
Take five minutes to get complete visibility into the performance, and start finding those needles, of your production applications with AppDynamics today.