In many IT environments today, nearly everything is shared and virtualized: pooled multiprotocol storage, virtualized CPU, memory and NICs, shared network resources for voice, video, storage and application traffic. This makes nearly everything interdependent. Interdependence can make monitoring, diagnosing, resolving and planning more difficult.
Monitoring is more complex because there are more and new things to monitor; some are physical and some are virtual. It can be more challenging to figure out what’s going wrong and traditional fixes often don’t work because of interactions between applications. Planning for future growth may be an additional challenge if vendor requirements are specified in physical standalone terms.
There is a hierarchy of tools to help IT Managers with these challenges: centralized logging & alerting tools, visualization tools that allow easier view of trends and analytic tools. Analytic tools use knowledge about applications, past and current resource behavior to intelligently alert, capture and present information.
Application performance is largely determined by three things: getting information to and from disk (storage), CPU/memory load of servers, getting information to and from servers (network). Appropriate analytic tools provide knowledge of applications and statistical measure of recent behavior that simplify monitoring, diagnosing, resolving and planning.