- Fix an issue where a query (and some of its tasks) continues to
consume CPU/memory on the coordinator and workers after the query fails.
- Fix a regression that cause the GC overhead and pauses to increase significantly when processing maps.
- Fix a memory tracking bug that causes the memory to be overestimated for
GROUP BY queries on
- Improve the performance of the
- Add support for casting from
- Disallow having a database in the
connection-url config property.
- Decrease planning time by fetching index metrics in parallel.
- Allow predicate pushdown for ObjectID.