3PAR System Reporter User’s Guide | System Reporter 2.8 |
Table
Metric | Description |
|
|
Queue Length | Queue length at the sample time. |
| Note that (unlike the other metrics above), the queue length is an |
| instantaneous measure at the sample time, not an average over the |
| sample interval. |
| Please note that due to the way that RCFC ports process data, the |
| Queue Length may not be a valid measure. |
|
|
Avg Busy % | The percentage of time that the object is busy (i.e., has at least one |
| outstanding IO operation). |
| Note that this is not the same thing as the percentage of available |
| object bandwidth that is being used. For example, a port shown as |
| 50% busy may have much less than 50% of its bandwidth |
| utilized.This is because the port may be idle for a substantial period |
| of time between the request and response. |
| Please note that for RCFC ports, the value will always be at 100% |
| since the read requests are always waiting in the queue to be |
| processed. |
|
|
For Histogram At Time and Histogram Versus Time report types, the PD, Port, LD and VLUN Performance reports show the read counts, write counts or total counts in various service time buckets. Depending on the Count control (see Count on page 5.32), these counts are either the accesses within the sample interval or the total count from the beginning (object creation or system reboot) until the current sample.
The CPU and VV Cache performance reports provide different metrics that are described separately in the following sections.
6.12 | Performance Reports |