Paul S. Randal
SQL Server DBAs the world over have a problem that never seems to go away: troubleshooting, with the vast majority of troubleshooting being done to find a performance problem of some kind. Even the most carefully architected and tested application system will experience changes over time that can lead to significant performance issues.
For example, the workload can change (such as the number of concurrent users, the queries being performed, and new month-end reports being run), the volume of data being processed may increase, the hardware on which the system is running may change (such as the number of processor cores, amount of server memory available, and I/O subsystem capacity), and new concurrent workloads might be introduced (such as transactional replication, database mirroring, and change data capture).
Read More