Starburst Galaxy's most valuable features include query federation, cluster management, and performance enhancements. Users highlight the ability to handle large jobs with right-sized clusters and support diverse data sources through a consistent interface. Cross-catalog querying, AWS Glue compatibility, and streamlined data connections boost user experience. Flexibility in connecting warehouses, compute-focused architecture, Iceberg support, autoscaling, and metadata management help reduce costs, optimize workload management, and facilitate integration with platforms like Kafka, S3 Iceberg, Snowflake, and Redshift.
- "The most fundamental feature is the query engine, which is much faster than any of the competitors; Starburst is able to finish most queries within 10 seconds, which is especially important for many non-technical employees."
- "Starburst has provided us with virtually guaranteed performance on complex queries across datasets that are in the tens of gigabytes which complete in seconds."
- "Starburst Galaxy serves as our primary SQL-based data processing engine, a strategic decision driven by its seamless integration with our AWS cloud infrastructure and its ability to deliver high performance with low-latency responses."
Users desire improvements in Starburst Galaxy's cluster startup time, wishing for faster spin-ups and enhanced control over cluster configuration. There is interest in better Tableau and dbt integration, with more advanced connectors and documentation. They express a need for improved monitoring interfaces, alerting integrations, and support for more connectors like Kinesis and Firehose. Additionally, there is a call for greater observability, enhanced impersonation functionality for access control, and a Terraform provider for infrastructure-as-code workflows.
- "I would like Starburst to leverage AI to improve usability. Data lakes are complicated and difficult for users to explore."
- "Multi-tenancy could be improved. In order to have multiple environments for SSO, we maintain multiple tenants that are connected to different AWS accounts via the Marketplace."
- "Cluster startup time is another pain point, typically 3 to 5 minutes, which is not the worst with proper planning but can be annoying for ad-hoc work."
