Why Accurate Performance Management Is Becoming Harder to Do

    June 21, 2016

    By Scott Adams

    As IT infrastructures become more complex and spread across various environments, it’s becoming harder to determine which systems are performing well and which ones are floundering.

    There’s no question that today, IT performance is more directly tied to the business than ever before. But at the same time, many IT functions that were previously handled in-house are now being outsourced to vendors. While this can provide a number of business advantages, it also makes it more difficult to control costs, align business and IT goals, and generally understand what factors are affecting your performance.

    Now more than ever, IT professionals must be able to quickly pinpoint inefficiencies within their infrastructures while remaining well within their departments’ allotted budget. Pressure is mounting on organizations to deliver better applications more quickly through strategies like DevOps, but this push for agility doesn’t make it acceptable to lose focus on the bottom line. After all, it does the business little good to cut delivery times by several months if the effort is regularly causing you to exceed your budget.

    We need to understand the new challenges surrounding this staple of IT maintenance and adopt tools that will empower us to take them on — a new era in IT calls for a new kind of performance monitoring tool.

    Untangling End-to-End Performance

    The art of performance monitoring has always involved untangling a series of connections between various software and hardware, including operating systems, data centers, and servers. As technology develops and new vendors enter the IT space, the likelihood that any one method for measuring performance will retain its accuracy diminishes significantly. The smallest upgrade to a given software might affect the precision of your monitoring tool, and you would never know it if you weren’t actively vetting it for inaccuracies.

    However, the biggest factor contributing to the growing complexity of performance monitoring is the rise of the public cloud. Public cloud providers like AWS and Microsoft Azure use their own data centers and decide how much performance information they want to release to consumers, as this SearchCloudComputing article explains. What’s more, the management system you use to look under the hood of your data center must integrate with the system used by your provider.

    When companies end up wasting their budget on poorly-performing applications, the cost benefits of the public cloud are often neutralized. For example, if your public cloud provider is only telling you that a given application is regularly meeting SLAs, you’d think it was performing perfectly — even if it was guzzling valuable resources in order to do so. Without a comprehensive, fully integrated view of your infrastructure, you might be wasting thousands on high-performance servers and software and not even know it.

    Finding an End-to-End Solution

    The primary goal of any performance management software should be to provide a view of your IT operations that is comprehensive — it accounts for the entirety of your infrastructure and any complications that arise from interactions between various environments and products.

    Some cloud vendors are responding to this demand with tools that offer a single-view look at both on-premises and cloud-based infrastructure. Again, these tools won’t necessarily be robust, and since they’re bespoke, it isn’t likely that they’ll give you the deep visibility into your specific operations that you’re looking for. It’s also unlikely that they’ll evolve in sync with the technology that they monitor, or remain accurate as your company shifts its strategy and infrastructure.

    Performance monitoring works best when it’s not only integrated and comprehensive, but completely customizable. As the SearchCloudComputing post argues, “if an organization customizes a cloud service, the visibility into its performance is often limited.” When shopping for an end-to-end performance management solution, organizations should be prioritizing an ability to customize the metrics they feel best measure performance. The more flexible your solution is, the more value you extract from your initial investment in it.

    As we’ve seen, the promise of technology like cloud computing can be deceiving if you don’t have visibility into how effectively you’re meeting SLAs and how much it’s costing you to do it. While IT focuses on goals that are more and more central to the business, it cannot lose sight of the fundamentals — ironically, they’re only getting more complicated as time goes on.

    (Image credit: Mick/flickr)