No alert mechanism to notify users when the memory guaranteed reservation allocation exceeds the available resources in VCD

Products

• VMware Cloud Director

Issue/Introduction

• When FLEX allocation is used with 100% memory guaranteed in an OVDC, all VMs will consume the full guaranteed memory. If memory runs out, new VMs won't power on.
• The user is unaware of potential resource allocation mismatches or non-compliance issues until they explicitly modify the reservation, rather than being proactively alerted when the reservation exceeds available resources or when the VM is non-compliant.

Environment

VMware Cloud Director (VCD) 10.x

Cause

This behavior is likely due to the system not continuously monitoring the reservation compliance unless a configuration change occurs.

Resolution

Resolution:

• Currently, there is no fix available for this issue. It is planned to be addressed in the next release.

Workaround:

• You can run an API call on each VM to check for compliance with the new memory allocation policy.
• Use the following API to check if a VM is compliant with the policy: 

POST https://<vcd-url>/api/vApp/vm-<id>/action/checkComputePolicyCompliance

• You will receive a prompt to make the VM compliant.
• Click "Make compliant", the compliance action will be successfully applied.

For large environments, writing an automation script will be the most efficient solution. To automate the process, follow these steps:

• Fetch the List of VMs:
• Query each VM to check for compliance by making the following API call:

POST https://<vcd-url>/api/vApp/vm-<id>/action/checkComputePolicyCompliance

• Check Compliance:
• For each VM, check the compliance status by looking for "checkComputePolicyCompliance" in the API response.
• If a VM is non-compliant, trigger the reapplyComputePolicy API to apply the policy:

POST https://<vcd-url>/api/vApp/vm-<vm-id>/action/reapplyComputePolicy

• After performing the POST request, respective VM users will start receiving the compliance alert.