After installation of an OS, companies detecting problems in end to end transactions. If I check how, for example, Windows™ has been installed, they tell me, next, next ,next. They did not look how to get the best performance out of the system without increasing resources.
The first thing you need to realize is the following: Prevention is better than cure; this can save your company a lot of money, it is cheaper to prevent than resolving every time.
Second part is > diagnose before it actually happened: moving form firefighting to proactive is the best thing you can do. Actually, with the tools like analytics you can predict what, when and why you can expect some events and take the necessary steps before they actually start to delaying the performance of your transactions, or when events are taking place, try to avoid them and reduce the number of problems with the right monitoring tools.
Last step: Resolve and prevent before it happens.
Create a checklist of problems and relation ships between problems, and think how to avoid them:
A lot of times when I do a health check about performance I find the following settings on a server:
Why do you need animate controls or shadows under a mouse pointer? This is the default, nice for a desktop but not for a server; use “adjust for best performance” on a server.
You can get a lot more capacity out of your hardware by changing this kind of parameters, by changing all of those parameters your transactions will go faster and you get more capacity out of your servers. Saving money, happy end users and more out of your existing hardware. ITPA has written an AI knowledge base to automate this kind of things.
24 january 2016