Have you ever experienced all hell breaking loose in the IT department? Likely, your answer is ‘yes.’
It happens all too often when there’s a breakdown in the IT infrastructure that causes an application to be sluggish or non-responsive. Then, the phone rings non-stop and help requests flood in. Users complain. They want immediate action so they can remain productive and keep customers happy.
IT administrators who do not have performance monitoring tools to guide them, or who only have rudimentary software, typically start scurrying down every conceivable rat hole to find and solve the problem. However, going strictly on instinct without data and metrics to guide them eats up their time; it leaves them discouraged and users angry.
That’s why you need an IT infrastructure performance monitoring tool that shines the light on any problem that’s disrupting the user experience.
There are many features to look for in a tool, but here are two essentials that help prevent the firefights and save you time. Not only do they spotlight the issues, but they will also help you avert serious problems before they occur.
Heart disease is known as a silent killer. You may not be aware that you have it until you feel tightness in your chest and, perhaps, experience dizziness. By then you may already be having a heart attack, and it could be too late to save your life.
The pain of the heart attack is similar to the alerts that many storage performance monitoring tools provide — they come too late. The alert lets you know that you’re out of storage space, CPU is pegged, or latency is running high. At this point, the problem is already affecting users and customers; you need all hands on deck to resolve it immediately.
To avoid 911 calls and preventable deaths, doctors and patients regularly monitor their patients’ vital signs….In the same way, you want a performance monitoring tool that warns you before a problem occurs so you can take preventive action….
To avoid 911 calls and preventable deaths, doctors and patients regularly monitor their patients’ vital signs, such as their blood pressure and pulse rates. If any of key numbers go beyond certain thresholds, doctors advise patients to take preventive action, such as changing their diet, exercising more and taking medications.
In the same way, you want a performance monitoring tool that warns you before a problem occurs so you can take preventive action and avoid slowdowns and downtime. It should capture trends, for example, in CPU and memory usage, IOPs, throughput, and latency. These are the vital signs of your IT infrastructure. When it spots a trend that could lead to a problem, it ought to let you know what the problem is and how long you have to resolve it. For example, “If this trend for CPU usage persists, it will be pegged in 13 days.”
The thresholds and associated warnings help you to prioritize where to take action first before users even know that an issue is brewing.
The boy who cried wolf when there was no cause for alarm did it so frequently that when a wolf was really stalking his sheep, the villagers failed to react to his calls.
You can run into a similar situation if you have a performance monitoring tool that’s always telling you there’s a problem. Often, there are spikes in activity that cause momentary slowdowns that users would not notice. Essentially, as each spike cries wolf, your performance monitoring tool runs the risk of being disregarded as white noise. After all, you have a choice of squandering your time by reacting to everything or ignoring the alerts because most of them are meaningless.
You need a tool that warns you only when you need to pay attention.
Thus, the best performance monitoring tools will eliminate one-time spikes in the data and let you know, for example, when there are several intervals in a row above a threshold that last five minutes or more. This feature enables you to focus on what matters most, the issues that affect the user’s experience.
So when looking for a performance monitoring tool for your IT infrastructure, seek one that provides predictive analytics, letting you know about potential problems before they occur. Also, make sure that you won’t be alerted to every single spike in the data, just those that require you to take action.