I have received a report on performance and capacity metrics for our server state - and as I am new to this area (not from a networking or infrastructure background), I am trying to gain some useful management information from them, to see if any warrant further investigation or suggest there are problems.
Some of the metrics focus on disc IOPS, writes/sec, reads/sec. All the servers in question are virtual machines with SAN based storage. It gives figures on disc IOPS in numbers, i.e. 4.89, at both an average (over an hour taken at 5 minute intervals), and a maximum found during an hour period, i.e. 12.26. The report lists the same for writes/sec, and reads/sec, also in number format, also taken at 5 minute intervals over an hour, also showing the average value and the maximum value. And the last columns in the report also show average disc read and write queue lengths, also shown as a decimal figure, ranging from 0.01 to 1.92 for writes, and 0.01 to 2.52 for read averages.
What exactly would pose an alarm on these values, i.e. high values or low values?
What would cause a problem (be that high or low values), and perhaps more importantly, what would need to be done to address this, i.e. is this an indication of insufficient hardware capacity for these servers?