NS0-701: Performance Monitoring and Troubleshooting

What this domain tests

Performance items ask you to gather the right counters, separate a capacity problem from a latency problem, and identify when QoS is the limit.

Core rules

• QoS policy groups define maximum throughput and maximum IOPS. A non-adaptive group applies its limit as configured; an adaptive group scales the limit with the used capacity of the workloads assigned to it.
• QoS can be assigned at the volume level or inherited from the SVM level, so always check both with qos policy-group show and volume show -fields qos-policy-group.
• qos statistics volume latency show breaks latency into network, cluster, data, and disk components, which tells you where to look next.
• Node-level checks include system node run -node <node> sysstat -x 1 for CPU and disk utilization, and statistics show -object volume for per-volume counters.
• ONTAP emits EMS events for thresholds such as volume full and high CPU; event log show is often the fastest first step.

Worked example

Users report slow NFS writes on one volume while other volumes are fine. qos statistics volume latency show shows the volume pinned at its policy group ceiling, and volume show confirms a 500 MB/s limit inherited from the SVM. Raising the policy group's max throughput restores performance without touching disks, which proves the bottleneck was QoS, not hardware.

Traps

Reading a single sysstat sample instead of watching a trend. Blaming disk latency when the QoS ceiling is the cause. Confusing a full volume with a slow volume - check percent-used before tuning anything.