The VM Rightsizing dashboard helps you adjust the VM size for optimal performance at the lowest cost. It covers both undersized and oversized scenarios. The dashboard looks at the long term trend and covers both undersized and oversized scenarios. This dashboard is designed for the Capacity and the Operations teams, as rightsizing a VM helps in the day-to-day performance.
How to Use the Dashboard
- Overall Analysis
The scoreboard provides a summary of the total undersized and oversized CPU and memory.
You can select either a data center or a cluster. In most cases, rightsizing analysis should be done at the cluster level as VMs typically do not move inter-cluster. The counters are displayed to provide better context. Focus on reclaiming VM capacity in a cluster that is low on capacity remaining.
The distribution charts that display rightsizing are automatically displayed. Other than the bar charts, the Undersized VM and Oversized VM tables list the actual VMs.
- VM Analysis
Select a VM to investigate further. The utilization is automatically displayed. VM usage reflects the amount of capacity consumed. This is based on the aggregate vCPU usage at five minute granularity, which provides a clear understanding of capacity used.
Rightsizing a VM can help improve performance for the VM as well as the cluster. Because of this, metrics that show performance bottlenecks, such as CPU Ready and CPU Run Queue, are provided to help you confirm how rightsizing may result in less contention and better performance.
Memory utilization is collected from the guest OS via VMTools. If guest OS metrics are not available, then the memory configured value is used instead. Rightsizing memory improves performance by reducing memory ballooning and contention. For example, VMs with over-provisioned memory are more likely to experience ballooning.
