• Share this post Keep up with us Summary Protect workloads against capacity errors: When your preferred VM type isn’t available, Databricks automatically falls back to compatible alternatives so clusters can still launch.Get Fleet-style flexibility on every cloud: Flexible node types bring automatic instance type fallback to Azure, GCP, and AWS. • Enjoy a simpler “1-click” workspace-wide activation, with clear visibility into acquired resources and optionally configurable fallback ordering.Reduce spend without sacrificing reliability: Prioritize discounted Spot instances when available, and fall back only when needed to maintain launch success. • Protect workloads against capacity errors: When your preferred VM type isn’t available, Databricks automatically falls back to compatible alternatives so clusters can still launch. • Get Fleet-style flexibility on every cloud: Flexible node types bring automatic instance type fallback to Azure, GCP, and AWS. • Enjoy a simpler “1-click” workspace-wide activation, with clear visibility into acquired resources and optionally configurable fallback ordering. • Reduce spend without sacrificing reliability: Prioritize discounted Spot instances when available, and fall back only when needed to maintain launch success.

Article Summaries:

  • Databricks has released “Flexible Node Types” as a generally available feature, enabling clusters to launch even when the preferred cloud instance type is unavailable. The system automatically falls back to compatible instance types, reducing capacity‑related launch failures that previously caused “stockout” errors on AWS, Azure, and GCP. For Spot‑with‑fallback clusters, the feature first seeks Spot capacity across the fallback list before switching to On‑Demand, boosting cost efficiency. Administrators can enable the feature with a single click in workspace settings, view acquired node types via the node_timeline table, and optionally define custom fallback orders through the API.

Sources: