Bug Fixes
- #6711 Fix heap use after free during redpanda shutdown by @mmaslankaprv in #6927
- #6809 Fix the search path for device info when executing
rpk redpanda tune disk_irq
by @r-vasquez in #7339 - Fix a bug that prevented redpanda from uploading the last batch in the log to cloud storage if timeboxed uploads were enabled and the batch contained exactly one message by @ztlpn in #7168
- Fix an issue where state machine snapshots could degrade performance under certain workloads (#6854) by @jcsp in #6983
- pandaproxy: Shut down consumers more gracefully during shutdown. by @BenPope in #7303
- pandaproxy: consumer fetch: More gracefully handle partition movement by @BenPope in #7303
- Fix automatic partition rebalancing when multiple nodes are added to the cluster before previous rebalancing finished (#7367) by @mmaslankaprv in #7389
Features
- Add POST v1/partitions/rebalance API to manually trigger partition rebalancing (#7367) by @mmaslankaprv in #7389
Improvements
- The rpk tune clocksource will display a better error message when executed from machines with unsupported architectures (#7442) by @r-vasquez in (#7443)
- archival: An ERROR log message on S3 upload failures is downgraded to WARN, as this situation has been observed to occur transiently in normal operation on AWS. (#7481) by @jcsp in #7482
- Change HTTP error code when leadership transfer can't proceed due to recovery from 504 to 503 (#7444) by @dlex in #7445
- cloud_storage: Improve tiered-storage reliability in presence of inconsistencies in metadata by @Lazin in #7508
- rpk will prevent decommissioning a node if its in maintenance mode. (#7474) by @r-vasquez in #7499
Full Changelog: v22.2.7...v22.2.8