Performance Monitors That Get Results

摘要

Monitoring IT service delivery is the endgame of all performance-management products. The primary benefits are monitoring a service's availability over time and pointing out possible causes of slow and failing performance. You can take the guesswork out of the process by periodically gathering snapshots of system-, network-, application-and client-performance capacity. A diagnostic record of what is "normal" can help you determine where to begin tackling a problem. Performance products gather a bewildering amount of information, from simple system availability to CPU percentage per Java process, from the response time of a particular HTML path on a per-user basis to a dissection of that total path time into network, server, database and client think time. These products will store, average and correlate over time all the performance metrics you choose, creating a mountain of data. To make sense of it all, you must know what performance information you need. In addition, you should understand the advantages and disadvantages of different measurement techniques to determine which performance tool will best suit your needs.