•
Discovery API:dataproc.v1↗
High Impact
Docs

Dataproc Introduces Flexible Instance Groups for HO Clusters

Developers can now leverage High Obtainability (HO) flexible instance groups for Dataproc clusters. The ClusterConfig schema gains three new fields: masterFlexibleConfig, workerFlexibleConfig, and secondaryWorkerFlexibleConfig. These fields reference the new FlexibleInstanceGroupConfig schema, which allows for granular specification of instance properties like bootDiskSizeGb, coreCount, and numInstances.

A notable addition within FlexibleInstanceGroupConfig is the hardwareClass enum, which simplifies instance type selection with options like STANDARD and PERFORMANCE, abstracting away specific machine types. This new configuration paradigm is designed for enhanced resilience and offers a single source of truth for desired cluster state. Additionally, the SessionTemplate schema now includes an optional displayName field for user-friendly identification.