The Secrets to Attentive Capacity Management
In today’s complex, heterogeneous environments, capacity managers need to lean on algorithms to predict usage. But never forget: it takes hard work and human savvy to run an efficient operation.
Reading about the status of capacity management today, it’s easy to get the sense that algorithms and predictive analytics will soon replace every critical IT management function (and with the current scarcity of talent, some deeply hope that that’s the case). And while it’s true that much of the increasingly difficult work of predicting server demands and allocating capacity can be automated, it doesn’t mean that IT professionals can just sit back and let the computers do the work.
Organizations need just the right combination of sophisticated tools and old-fashioned know-how to optimize their data capacity planning. Without people-oriented business goals, capacity management is rudderless. Before you can walk the fine line between overspending and overload, you need to learn these secrets of attentive capacity management.
To be sure, IT professionals are always separated from the IT infrastructures on which they operate by one degree; they depend on high-tech tools to compile comprehensive capacity data and render it intelligible and actionable. With any luck, that data will also come paired with insightful predictions, warnings, and recommendations. Strong tools for compiling and analyzing data are the foundation of any IT worker’s ability to do their job effectively.
But as a TechTarget article observes, it’s IT professionals who have to identify where these tools should be applied, to what degree, and to which objective — and that makes all the difference.
Before capacity management happens, workers have to carefully calibrate tools for their specific IT environments; when should thresholds trigger alarms for high utilization? Which KPIs are important? Will you monitor every application, virtual machine (VM), host, or cluster?
Moreover, when data begins to come in and algorithms start to identify patterns, IT workers have to decide which trends are salient, and why. Trends can be misleading: for example, in elastic, virtual environments, seemingly remarkable apps may only perform well because its resources are being over-allocated. Likewise, problems that build slowly can become unintentionally normalized and obscured by VMotion. The more IT workers are dedicated to poring over this data with vigor, the less likely it will be that trends are mistakenly validated.
This takes a constant evaluation of usage data, performance thresholds, future use models, and double checking to make sure that you’re operating in the right scope. Most important of all, however, is ensuring that IT operations are always supporting business goals. As TechTarget implores, don’t forget to “call in the suits.”
Many, if not most capacity management decisions (e.g. investing in increased bandwidth, migrating an app to the cloud, or confirming that you don’t require more capacity) largely depend on your desired business outcomes. For instance, if a server has 60% utilization, that may indicate that you need to simply reconfigure its hosting. But if you’re planning to expand an app’s service level, it might be wiser to instead reserve a cloud instance in the event that traffic spikes.
In other words, making the right call not only requires a close collaboration between IT professionals and their tools, but also between those tools and management. If capacity management systems can take IT problems and put them into a clear business context, it better enables IT workers to manage capacity attentively. At that point, optimization efforts actually begin to reap optimal value.
It’s unlikely that there will ever be a scenario where IT can kick back and let automation completely take over — they’ll always need to be there to guide capacity management tools in the right direction. For that to happen, business and IT need be attentive to one another’s needs.
(Main image credit: Eranger/Pixabay)