Vmware: definitions for metrics, properties and alerts

The VMware vRealize Operations reference guide provides an in-depth overview of metrics related to IT components’ performance and efficiency, such as virtual machines, hosts, and storage systems. This document is crucial for projects focusing on energy efficiency in IT, as it includes metrics that help monitor and optimise the power usage and overall resource efficiency of data centre operations. Key energy-related metrics include:

• Power Usage Metrics: These metrics track energy consumption (in Joules) and power usage (in Watts) across virtual machines, hosts and cluster compute resources, which is crucial for identifying high energy consumption and potential savings areas. The key power metrics proposed are:
  • Power|Energy (Joule)
  • Power|Power (Watt)
  • Power|PowerCap(Watt)

• Sustainability Metrics: These metrics are collected for a virtual machine, the host system, and cluster compute resources. 
  • Power|Total Energy (Wh)
  • Sustainability|CO2 Emission (Kg) (Cluster Compute Resource):
  • Carbon dioxide Calculated as power consumption* CO2 Emission rate Formula: CO2 Emission (Kg) = Sum(Host System(Power|Total Energy(Wh)))/ 1000 * 0.709
  • Sustainability|CO2 Emission before Virtualization (Kg) (Cluster Compute Resource): 
  • Carbon dioxide emissions before virtualization, assuming power consumption per physical server is 100W, reflecting a low-end hardware specification. Formula: CO2 Emission before Virtualisation (Kg) = Summary|Number of Running VMs * 0.1 * 0.709
  • Sustainability|CO2 Emission by Idle VMs (Kg) (Cluster Compute Resource)
  • Total carbon dioxide emission from all idle VMs. Calculated as CO2 emission rate * Power consumed by Idle VMs, where the rate is set at cluster property / 1000. Formula: CO2 Emission by Idle VMs (Kg) = Sum(VM(Power|Total Energy (Wh)), If VM(Summary|Reclaimable Idle = 1) * 0.709 OR CO2 Emission by Idle VMs (Kg) = Power Wasted by Idle VMs (Wh) * 0.709
  • Sustainability|Electricity Cost Savings (Cluster Compute Resource)
  • Estimated cost savings by virtualising workloads. Calculated from the difference between power consumption before and after virtualisation. The electricity cost is defined at the cluster custom property. Formula: Electricity Cost Savings = (Summary| Number of Running VMs * 0.1 – Sum(Host System(Power|Total Energy(Wh)))/1000) * 0.108 OR Daily Electricity Cost Savings = (Summary|Number of Running VMs * 0.1 – Power usage (KWh)) * 0.108
  • Sustainability|Power usage (KWh) (Cluster Compute Resource)
  • Power usage calculated from all hosts in KWh. Formula: Power usage (KWh) = Sum(Host System(Power|Total Energy(Wh))/ 1000
  • Sustainability|Power usage per GHz (Wh) (Cluster Compute Resource)
  • Power usage efficiency. Calculated as power consumption over total GHz. Formula: Power usage per GHz (Wh) = Sum(Host System(Power|Total Energy(Wh))/CPU|Usage (MHz)/1000
  • Sustainability|Power Wasted by Idle VMs (Wh)
  • Sum of electricity power used by all VMs classified as idle by the system. Formula: Power Wasted by Idle VMs (Wh) = Sum(VM(Power|Total Energy (Wh)), If VM(Summary|Reclaimable Idle = 1)
  • Sustainability|Trees to Offset Idle VMs CO2 Emission (Cluster Compute Resource)
  • Number of standard trees required to compensate CO2 emission of all Idle VMs. Based on 36.4 pounds of carbon per tree. Formula: Trees to Offset Idle VMs CO2 Emission = Sum(VM(Power|Total Energy (Wh)), If VM(Summary| Reclaimable Idle = 1)/1000 * 0.709 / 16.511 OR Trees to Offset Idle VMs CO2 Emission = Power Wasted by Idle VMs (Wh)/1000 * 0.709 / 16.511

By leveraging the detailed metrics outlined in the document, IT administrators can gain valuable insights into their systems’ energy consumption patterns and take proactive measures to enhance energy efficiency. These measures include rightsizing resources, optimising workload placement, and using predictive analytics to prevent performance bottlenecks that might lead to increased power usage.