In a recent update, Elasticsearch introduced several notable improvements and changes to enhance user experience and address existing issues. These updates include better query cancellation, which now propagates cancellation to all nodes and child tasks to prevent unnecessary CPU and I/O usage, although some operations like terms dictionary expansion and complex BKD reads still need more efficient handling. Mapping validation has been improved to catch misconfigurations earlier in the process, especially during index creation, providing users with a more seamless experience. The migration from Joda to Java time revealed the need for better backward compatibility, prompting Elasticsearch to restore support for pre-7.0 indices and improve migration documentation. Lucene's integration now includes the updated Snowball stemmer, offering support for additional languages such as Hindi and Tamil. Integrity checks via Lucene's CheckIndex tool are being enhanced to catch bugs earlier, and a proposal for a more flexible Korean dictionary mechanism is under consideration. The update also introduces breaking changes like disallowing duplicate percentile values in percentile aggregations and boosting the default scroll size in the Elasticsearch Hadoop Plugin.