Guy Warren, ITRS
As IT support professionals, we are always concerned about failure of our application estates, or severe degradation in performance, due to 'hotspots'. There are parts of the IT estate which become overloaded at periods of peak business load. If everything were uniform and consistent through time, this would be relatively easy to detect and address, but IT estates aren't. Business volumes are unpredictable, and are often affected by external factors in the market - think CHF/EUR uncoupling, Grexit or the Yuan revaluation. And the IT estate is ever-changing with software releases and other activities running on servers which can impact their performance, such as garbage collection.
This means it is necessary to understand what the IT estate is doing at all times, in real-time, to see if hotspots are growing or if they have changed since a period of known performance. Is the current hotspot activity unusual compared to expected business volumes? Have the recent software releases or other changes affected the estate's performance?
Real-time monitoring of the server's CPU, memory and processes across a large estate is very challenging, but spotting changes in behaviour was previously next to impossible. As businesses look to modernise their IT estates and optimise performance while reducing costs, it becomes a necessity to capture machine data from all servers in real-time, and to compare it with previous periods to see if the estate is behaving 'normally' or within given parameters.
So what about cold spots? The need to have sufficient capacity at all times in all circumstances means that there is little attention paid to consistent, surplus capacity. Parts of the estate never use their full capacity and can be a waste of a valuable resource, but identifying these servers has historically been time-consuming and difficult to do, and therefore not a priority. With smarter tools becoming increasingly available, it is important to identify these cold spots, and look at options to reduce spend on these parts of the estate through down-sizing or virtualisation.
Hotspots or cold spots - the modern IT support professional needs real-time analytics tools to help manage their estate, prevent problems or outages, and save valuable resources.