The disk can be the first bottleneck for your virtual environments; for this reason, it's very important to understand how much is used and how it is used, both for the capacity (but this could be quite easy) and for the expected performance. Using a simple input/output operations per second (IOPS) approach could work in several cases, but does not appropriately qualify the type of I/O and cannot always help in the storage sizing and design. Storage will be discussed in detail in Chapter 7, Advanced Storage Management.
For the disk, the following table lists some of the common metrics:
|
Metric |
Description |
Why it's important |
|
% disk time |
Percentage of elapsed time that the selected disk drive was busy servicing read or write requests. |
Could be useful to identify workload that is potentially I/O bound. |
|
Disk latency (average and peaks) |
Observed latency of the disks. Warning zone of latency is typical around 10-20 ms. |
High latency could mean disks or other storage bottlenecks. |
|
Disk bytes/sec |
Represent the rate of the bytes per second that are transferred to or from the disk during write or read operations. |
This provides information about the throughput of the disk system and how busy it is. |