Concepts
Workspaces
Workspaces are isolated, top-level environments that group users, projects, resources, billing, and policies for a team or business unit. They provide centralized configuration (network, storage, integrations) and membership controls so you can separate development, staging, and production boundaries cleanly.
Clusters
Clusters supply compute (CPU, RAM, autoscaling) for interactive analysis and batch jobs; use ephemeral clusters for ad-hoc work and persistent clusters for production. Manage lifecycle, sizing, and autoscaling to balance performance, cost, and workload isolation, and secure clusters with network and IAM policies.
Catalogs (Metastores)
Catalogs store table and schema metadata (locations, partitions, formats) so compute engines can discover datasets reliably across object stores. Use centralized, Hive-compatible metastores or external catalogs to enable schema sharing, governance, and consistent dataset discovery.
Access Control
Access control (RBAC/ACLs) enforces who can view or modify workspaces, clusters, catalogs, jobs, and notebooks, and integrates with SSO/IDPs for centralized identity. Apply least-privilege roles, group-based policies, resource-level ACLs, and audit logging to meet security and compliance needs.
Jobs
Jobs are reproducible, scheduled or event-driven pipelines that run code with parameters on a specified cluster profile; include retries, alerts, and artifact/version tracking. Use jobs to operationalize ETL, ML training, and production workflows, and monitor runs with logs and metrics for observability.
Notebooks
Notebooks combine code, visualizations, and narrative for exploratory data analysis and prototyping; attach them to clusters for interactive execution. For reproducibility, pin environment specs and convert stable analyses into parameterized jobs or scripts for production scheduling.
The recommended resolution for Yeedu UI is 1920 x 1080.