VKS Health Status Abnormal: "POST /pbm: 503 Service Unavailable" Error Due to vmware-sps Service Failure

Products

• VMware vSphere Kubernetes Service

Issue/Introduction

In a VMware vSphere Kubernetes Service (VKS) or Supervisor cluster environment, you observe the following symptoms:

• The VKS environment or Supervisor service reports an unhealthy state.
• Health checks or storage-related tasks fail with the error: POST "/pbm": 503 Service Unavailable.
• The vmware-sps (Profile-Driven Storage) service on the vCenter Server Appliance (VCSA) is stopped or not responding.

Environment

VMware vSphere Kubernetes Service (VKS)

VMware vCenter Server 8.x

Cause

The issue is caused by the vmware-sps service entering an unresponsive state or crashing following a Java Heap Out-of-Memory (OOM) event.

This causes the SPBM API endpoint (/pbm) to become unavailable, returning an HTTP 503 error to callers like VKS.

Resolution

To restore connectivity and service health, follow these steps:

1. Log in to the vCenter Server Appliance via SSH using root credentials.
2. Check Service Status: Run: service-control --status vmware-sps
3. Restart the Service: Run: service-control --restart vmware-sps

Additional Information

Review the following logs for evidence of memory pressure or recurring heap dumps: /var/log/vmware/vmware-sps/java_error###.log

If the issue recurs frequently, consider increasing the Java heap memory allocation for the SPS service according to your environment's scale.
For more information, run cloudvm-ram-size -h in a vCenter Server Appliance.