Understanding the Summary| IsGreenForPlacement metric

Products

• VCF Operations/Automation (formerly VMware Aria Suite)
• VCF Operations

Issue/Introduction

• Memory Stats not correct from Aria Operations (AOPs)
• Used the API call to get the values of the memory|is_workload_non_green metric - the value was 0
• The cluster summary in vSphere UI shows that there is NO available Memory in the cluster
• The API call returns data that seems incorrect

Environment

• VCF Operations 9.x
• Aria Operations 8.x

Cause

• The vSphere Console displays Consumed Memory.
• VCF Operations provides placement recommendations using the Workload Memory.

Resolution

Verify the available Workload Memory in the Cluster

1. Log in to the Operations UI and search for the Cluster object

2. Click on the Metrics tab and expand Metrics > Memory

3. Double click Workload (%)

4. Verify the latest value calculated is less than the threshold defined in the policy (80% by default)

Additional details for the relevant metrics

From the vSphere Client

• Consumed Memory %

  • Metric is the amount of memory used on the ESXi host in total. This includes memory used by the VMKernel, other vSphere Services and Management agents, plus the total consumed memory for all the running virtual machines. The latter part is however not an actual sum of what each of the guest operating systems in these virtual machines are consuming internally as the ESXi host is able to employ memory sharing technique. With memory sharing, a workload running on a virtual machine often consumes less memory than it might when running on physical machines. As a result, higher levels of overcommitment can be supported efficiently.

  • References
    • Memory Virtualization with vSphere
    • vSphere Resource Management
    • KB 326248 Understanding the Memory Active and Memory Usage indicators on the Performance tab

From VCF Operations

• Memory | Workload %

  • Calculated as Demand / (Usable Capacity after HA and Buffer) * 100

  • Note that Workload is calculated based on resource demand rather than consumed or used memory.

• Summary | Is memory non Green

  • Returns 1 if memory workload >= workload threshold (80% by default)

  • Returns 0 if workload < workload threshold (80% by default)

  • Value 0 is positive for placement and value 1 is negative for placement

• Summary | Is Green For Placement

  • With this metric, you are able to monitor if placement on the cluster can be considered at all.

  • The metric shows if either CPU or MEM workloads have gone over the set thresholds (80% by default)

  • The metric value returns 1 if both CPU and MEM workload percentages are below the set thresholds (is green), which means that placement can be considered.

  • The metric value returns 0 if one of the CPU or MEM workloads is over the threshold, which means that no placement can be considered.

  • The metric is based on Memory | Is Workload non Green and CPU | Is Workload non Green metrics’ negations

    • IsGreenForPlacement = NOT(Memory|Is Workload non Green) && NOT(CPU|Is Workload non Green)

• Capacity Remaining Percentage (%)

  • Based on the capacity remaining for memory (metric Capacity Remaining (KB)) and capacity remaining for CPU (Capacity Remaining (MHz))

  • Is the percentage of the lower point of capacity remaining for memory and CPU.

  • Reference How Does VMware Aria Operations Calculate and Forecast Capacity

Additional Information

Operations can provide additional workload placement information to Automation using the native integration between them. See the following documentation for more.

• Integrating Data Sources with VMware Aria Operations > VMware Aria Automation
• Workload Placement
• Advanced Workload Placement for Allocation Model