Setup
Iceberg tables on S3, queried via StarRocks (views with joins/aggregations)
Catalog: Polaris
Tableau Cloud with Tableau Bridge for extracts
Questions
1. Connector Recommendation
For extracting from StarRocks views, which is preferred:
MySQL connector, or
StarRocks connector?
Are there known limitations for large extracts (e.g., timeouts, fetch size, data type mapping, SQL pushdown)?
2. Extract Reliability (Bridge)
Best practices to ensure stable large extract refreshes, specifically around:
Long-running query timeouts
Bridge memory usage
StarRocks concurrency limits
Iceberg metadata latency
#Tableau Cloud
Connector Recommendation
- Use the StarRocks connector if available. It's better optimized for StarRocks, so you'll get proper SQL pushdown and more reliable performance.
- The MySQL connector can work, but may struggle with complex views, large extracts, and data type handling.
Extract Reliability (Bridge)
To avoid slow or failed refreshes:
- Run ANALYZE / keep stats updated for better query planning.
- Prefer incremental extracts over full refresh.
- Increase query/connection timeouts.
- Ensure Bridge machine has sufficient memory & CPU.
- Limit concurrent queries in StarRocks.
- Optimize Iceberg tables (compact files, reduce metadata load).
- Simplify heavy views (joins/aggregations).
The issue is usually not the connector alone, but how large queries are executed - use the StarRocks connector and optimize queries, resources, and extract strategy for stable performance.
Hope you find this helpful !!!! 😊
Thanks and Regards,
Pooja Bhattacharya
