Users are unable to access the VCF Automation portal due to multiple pods remaining in a pending state.

Products

• VCF Automation

Issue/Introduction

• The VCF Automation browser portal is inaccessible, and kubectl get pods shows several pods in "Pending" status. 
• VMSP node is Ready without disk or memory pressure issue
• default-scheduler failed to scheduling these pods due to "too many pods" 

Environment

VCF Automation 9.1

Cause

An excessive number of support-bundle-cluster-info-dump jobs consumed all available node resources, preventing functional pods from scheduling.

Resolution

1. Identify and delete all hanging support-bundle-cluster-info-dump jobs in the vmsp-platform namespace:

kubectl get jobs -n vmsp-platform | grep support-bundle-cluster-info-dump | awk '{print $1}' | xargs -I {} kubectl delete job {} -n vmsp-platform

2.Suspend the support-bundle-cluster-info-dump cronjob to prevent recurring resource exhaustion.

kubectl patch cronjob support-bundle-cluster-info-dump -n vmsp-platform -p '{"spec" : {"suspend" : true }}'

3.Verify that all pending pods transition to the "Running" state and validate portal accessibility.

kubectl get pods -n prelude | grep -vE 'Running|Completed'