Elevated CPU, IOPS, and OIO during Dynamic Threshold Calculation in VMware Aria Operations
search cancel

Elevated CPU, IOPS, and OIO during Dynamic Threshold Calculation in VMware Aria Operations

book

Article ID: 445637

calendar_today

Updated On:

Products

VMware Aria Operations (formerly vRealize Operations) 8.x VCF Operations

Issue/Introduction

Symptoms

You observe the following performance metrics in VMware Aria Operations daily between a certain time range.

  • Elevated CPU utilization.
  • Increased Input/Output Operations Per Second (IOPS).
  • High Outstanding I/O (OIO).

There is no reported service degradation or abnormality in data collection during this specific window.

Environment

VMware Aria Operations

Cause

These resource utilization spikes are an expected behavior of the VMware Aria Operations analytics engine. The vcops-analytics service initiates the Dynamic Threshold (DT) Calculation process daily. During this window, the engine executes intensive statistical computations against the telemetry data collected over the previous 24 hours. This process naturally drives peak disk I/O and CPU commit on the analytics nodes.

Resolution

No corrective software action is required as this represents healthy operational parameters for the analytics engine. To mitigate potential secondary impacts on your storage subsystem, you should apply the following optimizations:

  1. Validate Storage Latency: Ensure the underlying datastore sustains disk latency below 10ms during the DT calculation window to prevent localized I/O bottlenecks.
  2. Manage Virtual Machine Snapshots: Verify that no active hypervisor-level virtual machine snapshots exist on the Aria Operations analytic nodes. Active snapshots exponentially amplify I/O penalties and latency during high-OIO events.
  3. Optimize the Calculation Schedule: If the current execution window conflicts with infrastructure backups or other batch processing, you can modify the DT calculation schedule via administrative policy to an isolated off-peak maintenance window.
  4. Review Node Sizing: If CPU peaks consistently impact stability, evaluate increasing the node size (e.g., from Medium to Large) based on the sizing guidelines in the product documentation.