redpanda-data/redpanda v25.2.2 on GitHub

Features

A bug has been fixed that previously resulted in segment upload failures to take longer to be retried. by @andrwng in #27056
Fixes node wise recovery issue leading to stuck reconfiguration when local data on the replicas is more up to date than data in tiered storage by @mmaslankaprv in #27224
Redpanda now serializes Iceberg manifest lists differently, to allow certain engines (e.g. AWS Redshift) to query the Iceberg tables when using the empty partition spec. by @andrwng in #27206
#27317 Fixes a bug in which adjacent merge compaction could mutate a segment without advancing the segment's generation_id, which could potentially lead to transient upload failures in tiered storage. by @WillemKauf in #27322
logs_since flag when requesting debug bundle now handles timestamps provided in ISO8601 format. by @c-julin in #27210
repair initial leadership appointment machinery by @joe-redpanda in #27296
rpk: fixes a bug where describe_configs and alter_configs were not correctly parsed via rpk security acl by @r-vasquez in #27215

Improves frequent timeouts in cloud storage clients. by @oleiman in #27263
Makes admin redirects to leader a bit more flexible - allowing redirection of requests originally directed at a headless k8s service. by @oleiman in #27275
No longer generating audit events for internal Redpanda services by @michael-redpanda in #27031
Prevent redpanda from occasionally hanging when the application is exited soon after starting. by @WillemKauf in #27069
Reduces the fragmentation caused by ftruncate() calls on segments being rewritten as a part of compaction by providing a better estimate for the size required from initial fallocate() calls in the storage subsystem by @WillemKauf in #27038
#27266 rpk: lower memory consumption when running rpk debug bundle. by @r-vasquez in #27283

Full Changelog: v25.2.1...v25.2.2