Using weekly and monthly performance trends to identify performance issues

Identifying performance trends can assist you in identifying whether the cluster is being overused or underused by analyzing volume latency. You can use similar steps to identify CPU, network, or other system bottlenecks.

Procedure

  1. Locate the volume that you suspect is being underused or overused.
  2. On the Volume Details tab, click 30 d to display the historical data.
  3. In the Break down data by drop-down menu, select Latency, and then click Submit.
  4. Deselect Aggregate in the cluster components comparison chart, and then compare the cluster latency with the volume latency chart.
  5. Select Aggregate and deselect all other components in the cluster components comparison chart, and then compare the aggregate latency with the volume latency chart.
  6. Compare the reads/writes latency chart to the volume latency chart.
  7. Determine whether client application loads have caused a workload contention and rebalance workloads as needed.
  8. Determine whether the aggregate is overused and causing contention and rebalance workloads as needed.