Skip to main content

​
Stability Improvements 💪

  • Eagerly abort an index build when it will not fit on disk, failing fast instead of exhausting disk space.
  • Cap index build worker memory at 1GB to reduce peak memory usage during index creation.
  • Bound sequential scan memory usage by spilling to a BufFile.
  • Use the correct cardinality estimate when routing GROUP BY between Tantivy and DataFusion.
  • Return an error instead of silently returning incomplete results when Tantivy truncates results.
The full changelog is available on the GitHub Release.