Production performance metrics

Table A-1 describes the current production performance metrics to measure using operating system monitors.

**Table A-1: Metrics for production system**
Metric	What to measure
CPU	The average and maximum CPU utilization (aggregate and per CPU on SMP servers) per “time window” (online, batch, and so on) per server.
Disk I/O	Number of I/Os per second per disk and controller, and I/O queue lengths per “time window” per server. Total I/Os, reads, and writes per second per Sybase device per “time window” per server.
Concurrency	Determine the average lock contention. You can use sp_who to determine the processes currently running, and sp_lock to find out which current processes hold locks.
Network I/O	The packets per second per NIC per “time window” per server. The TDS packets (“sent from” and “received by”) per “time window” per server.
Memory	Determine the paging/swapping rates per second per “time window” per server. Also determine the data and procedure cache hit rates per “time window” per server.