spot
In-depth articles from the OrchesTerra engineering team.
Kubernetes
How Karpenter decides to replace nodes, how disruption budgets interact with PDBs, and the settings that stop consolidation from evicting production at the worst time.
2026-09-08 · 4 min read
Kubernetes
How to put real workload on Spot / preemptible nodes without turning every interruption into an incident — taints, PDBs, grace periods, and mixed on-demand capacity.
2026-09-01 · 5 min read