Performance, Monitoring, and Troubleshooting
This domain tests how to measure ONTAP performance, interpret QoS, and isolate causes of slow client access.
Performance concepts
ONTAP reports latency, IOPS, throughput, and utilization. QoS policies set maximum throughput and can be adaptive by scaling with used capacity. A volume or LUN can be throttled by its own QoS policy even when the node has spare capacity. Performance capacity used shows how much of a resource is consumed relative to its capability.
Monitoring tools
System Manager shows health and performance dashboards. CLI commands such as statistics show, qos statistics volume performance show, event log show, and autosupport status provide detail. Active IQ analyzes AutoSupport data for risks and best practices. EMS events should be forwarded to a monitoring system.
Worked example
Users report high latency on one volume while other volumes are responsive. Check the volume QoS policy first: if max throughput is set low, the volume is throttled regardless of node health. Then check aggregate utilization, node CPU, LIF errors, and client-side latency. Comparing qos statistics volume performance show output against node statistics separates a policy limit from a resource bottleneck.
Traps
Blaming disks before checking QoS; ignoring client-side or network latency; reading one latency sample without a baseline; and missing EMS events that indicate a failing disk or LIF.