Hi,
Our GKE Autopilot cluster (PII Removed by Staff) in asia-south1 is stuck
in RECONCILING state because SSD quota increase requests are being
auto-denied within seconds.
Project: (PII Removed by Staff)
Region: asia-south1
Current SSD usage: 200GB / 250GB (80% full)
Requested: 350GB - auto-denied twice immediately
The cluster autoscaler keeps trying to add new nodes but SSD quota
blocks it. This is affecting our staging environment with 17+
microservices.
How can we get a manual review for this quota increase?
Thank you
06/06/2026
Update: Found the root cause.
Each node is provisioning ~94GB ephemeral storage.
2 nodes × 94GB = 188GB + overhead = ~200GB used.
When autoscaler tries to add a 3rd node, it needs ~94GB more
but quota is only 250GB total.
Is there a way to request smaller node disk size, or get
the quota increased from 250GB to 300GB?
kubectl describe node shows:
ephemeral-storage: 98831908Ki (~94GB) per node
ephemeral-storage used: Node1=5Gi(11%), Node2=17Gi(38%)
Actual usage is very low — only quota is the blocker.