Capacity and Performance Metric Behavior for Clusters with Hosts in Maintenance Mode in VMware Aria Operations

Products

• VCF Operations/Automation (formerly VMware Aria Suite)

Issue/Introduction

In VMware Aria Operations 8.18.x, an administrator may observe that the "Usable Capacity" metric for a cluster does not immediately decrease when an ESXi host is placed into Maintenance Mode (MM). This behavior can lead to discrepancies between the resources reported in vCenter Server and the capacity analytics displayed within the Aria Operations UI.

Environment

VMware Aria Operations 8.18.x

Cause

The underlying cause involves how the Aria Operations Capacity Engine aggregates hardware resources.

By default, the environment may still include the hardware of nodes in Maintenance Mode in the "Total Capacity" metric because the objects still exist in the inventory.

While "Usable Capacity" is intended to exclude MM hosts, delays in the 5-minute polling interval or specific Policy configurations can result in the metrics remaining static until the host is either powered off or the collection cycle completes.

Resolution

To ensure capacity metrics accurately reflect the removal of resources during Maintenance Mode, the following should be considered:

1. Collection Interval: Verify that at least one full collection cycle (default 5 minutes) has elapsed since the host entered Maintenance Mode in vCenter Server.
2. Capacity Engine Policy:
• Navigate to Configure > Policies.
• Review the active policy applied to the Cluster object.
• Ensure that the settings for "Capacity Attributes" are configured to exclude non-active hosts if specific overrides are in place.
3. Total vs. Usable Capacity:
• Recognize that "Total Capacity" may still reflect the physical hardware present in the inventory.
• "Usable Capacity" is the primary metric that should decrease. If it remains unchanged despite the host being in MM, verify if the host is still being marked as "HA Active" within the Aria Operations object properties.
4. Hardware State: In some scenarios, powering off the host will force the Capacity Engine to immediately recalculate and reduce the "Usable Capacity" as the host is no longer reachable by the collector.

Additional Information

For further details on how capacity is calculated, refer to the - How Does VMware Aria Operations Calculate and Forecast Capacity