To determine if the virtual machine performance problem is temporary or indicates a more serious problem in the storage network, you create metric charts to identify the root cause of the problem.

You used the Storage Troubleshooting dashboard to find an object networked to the problematic virtual machine that is not operating at an optimum. Now you want to investigate the performance of that object over a period of time.

1

If you are on the Storage Troubleshooting dashboard with the problematic virtual machine selected, select a highlighted object that appears with a red, orange, or yellow condition.

2

Double-click the object to access collected metrics.

The left pane displays the object name and the related objects. The main pane displays the Summary tab.

3

Click the Troubleshooting tab, and click the All Metrics tab.

Throughput, latency, and error metrics are the most useful troubleshooting metrics.

a

Select the error statistics and double-click an entry to determine if errors occurred over time. Errors can indicate when the component had trouble accessing the network.

b

Select the derived statistics and double-click the use to see a historical graph of the component use over time. Slow throughput can occur during times of high workload.

c

Select the derived statistics and double-click the latency to see a historical graph of the time required to process IO requests. High latency can indicate that the component had trouble receiving data from the network.

If the error statistics show no problems and derived statistics show a cyclical use pattern on the component, you can call the user and explain that the problem might be temporary and when IO traffic throughout the network drops, the virtual machine will become more responsive.

If the error statistics show no problems and derived statistics show consistently low use of the component but the latency is suddenly high, another component on the network might be causing connectivity to time out. Check for a faulty switch or other faulty component on the network.