Tag See all →

#cluster-autoscaler

2 stories

FinOps & Cloud Cost Bug Report

Idle GPU nodes billed all weekend: why the cluster autoscaler would not scale down

Ryan Pham · Jan 14, 2026 · 6 min read

Our GPU node pool stayed at eight nodes from Friday night to Monday morning with almost no GPU usage. The cluster autoscaler refused to remove them: notebook pods with local storage, a bare debug pod and a kube-system pod without a PDB. Fixing each blocker cut GPU spend by 38 percent.

Banner: Kubernetes pods waiting in a queue in front of nodes marked with locks and full gauges

DevOps & CI/CD Bug Report

Kubernetes Pods stuck Pending: Insufficient cpu, untolerated taints and node affinity

John Nguyen · Jul 20, 2025 · 9 min read

A routine rollout left 14 Pods Pending with Insufficient cpu while dashboards showed nodes at 30 percent CPU. The scheduler counts requests, not usage, and a new tainted node pool hid the free capacity. Here is how we read the FailedScheduling message and fixed it.