Organizations exploring monitoring solutions should do the following: work with a vendor that can provide consulting about tools, processes and people, as well as help them evaluate representative solutions from all the options above; define clear quality and financial objectives and use them to evaluate the options above; keep in mind that the goal is to improve service and application availability not to monitor boxes, as monitoring hardware is not enough; and consider integrating monitoring vertically with other operational functions such as asset, incident and configuration management, capacity planning and security and patch management.
"In the beginning, the monitoring infrastructure was the users. Troubleshooting was done by primitive tools and elimination. The network was the usual suspect for any degradation. Mean time to repair was quite long, but it was acceptable. Not anymore. The network is now critical for human as well as machine communication. Problems should really be detected and resolved before users notice them and even before they become problems. Time to identify and engage the proper resources should be zero, and some problems now can be resolved automatically in seconds."