redpanda-data/redpanda v23.1.19 on GitHub

Features

#13534 crash_loop_limit now defaults to 5. If a broker uncleanly shutdowns for 5 times back to back, it is considered to be in a crash loop mode and Redpanda refuses to start up and may need a manual intervention. This enforcement is disabled in developer mode and rpk's dev-container mode. by @bharathv in #13686

Topic manifests now serialize size values as unsigned instead of signed. Previously negative values could result in the failure to recreate the topic as a recovery topic. by @andrwng in #13868
#13393 fixed very rare error when consumer group recovery was incomplete by @mmaslankaprv in #13728
#13541 rpk: prevent panic when using a directory instead of a filename in rpk redpanda tune --output-script by @r-vasquez in #13546
#13639 Fixes a case when a control batch is the last user batch in a segment and the log and is being compacted away. This gave a perception of consumption hang from a client POV but in reality there is nothing to consume after that point. Not discarding such batches lets the consumer offset make progress and reach LSO. by @bharathv in #13687
#13667 #12677 #13054 fixed rare but possible deadlock that might lead to Redpanda crash when moving partitions back and forth between nodes. by @mmaslankaprv in #13705
#13667 #12677 #13054 fixes reporting maintenance mode status for nodes which are offline by @mmaslankaprv in #13705
#13831 fixes a use_after_free due to interaction with iterators and suspension points. by @andijcr in #13963
#13846 Fixes an issue where autocreated topics would not reflect cluster-level cloud storage config by @oleiman in #13852
#13956 Schema Registry: GET /subjects no longer returns a subject that has no versions or only deleted versions (unless ?deleted=true) by @BenPope in #13957

#13435 The broker now logs errored Kafka API responses at the DEBUG level. by @NyaliaLui in #13436
#13721 Export per handler produce/fetch metrics on the public_metrics endpoint by @StephanDollberg in #13730
#13882 redpanda/kafka: Eliminate some copies in kafka responses (in particular, metadata) by @BenPope in #13897
#13958 Schema Registry: Improve error message for invalid compatibility level by @BenPope in #13959
#13968 kafka/client logs now include the client identifier (e.g., schema_registry_client, pandaproxy_client) by @BenPope in #14074
#14140 #14104 Kafka: Avoid very large contiguous allocations during fetch. by @BenPope in #14141

Full Changelog: v23.1.18...v23.1.19