redpanda

Features

• #14381 Schema Registry: Add DELETE /config/{subject} endpoint by @oleiman in #14383

Bug Fixes

• #12610 Fix a bug that could lead to crashes with double registration of group assertion after trying to create a lot of topics in a single CreateTopics request when some of these topics could not be allocated (e.g. due to bumping into partition count limits). by @ztlpn in #14465
• #13442 Fix a bug where the alter configs API would accept even-valued replication factors by @oleiman in #14338
• #14418 Fix OffsetForLeaderEpoch for read-replicas by @Lazin in #14421
• PR #14175 [v23.1.x] cloud_storage: Handle nested benign exception by @abhijat
• PR #14187 [v23.1.x] s3_client: Check if response contains error by @Lazin
• PR #14206 [v23.1.x] fix CDT broken tests by @gousteris
• PR #14306 [v23.1.x] cloud_roles: redact fields from request string representation by @abhijat

Full Changelog: https://github.com/redpanda-data/redpanda/compare/v23.1.19...v23.1.20

Features

• #14380 Schema Registry: Add DELETE /config/{subject} endpoint by @oleiman in #14382

Bug Fixes

• #14417 Fix OffsetForLeaderEpoch for read-replicas by @Lazin in #14419
• #14437 #14463 Do not retry waiting for tx_id hosting partition when current node is not leader by @mmaslankaprv in #14448
• #14399 [v23.2.x] build(deps): bump github.com/onsi/gomega from 1.28.0 to 1.28.1 in /src/go/k8s by @dependabot[bot]
• #14448 [v23.2.x] [backport] tx: add more trace logging in init_producer_id code path by @bharathv
• #14457 [v23.2.x] build(deps): bump github.com/onsi/gomega from 1.28.1 to 1.29.0 in /src/go/k8s by @dependabot[bot]

Full Changelog: https://github.com/redpanda-data/redpanda/compare/v23.2.13...v23.2.14

Discussion: https://github.com/redpanda-data/redpanda/discussions/14493

Bug Fixes

• #12610 Fix a bug that could lead to crashes with double registration of group assertion after trying to create a lot of topics in a single CreateTopics request when some of these topics could not be allocated (e.g. due to bumping into partition count limits). by @ztlpn in #14466
• PR #13818 [v22.3.x] rpk: run go tests in gh actions by @gousteris
• PR #13939 [v22.3.x] only run PR checks when needed by @joejulian
• PR #14207 [v22.3.x] tests: remove 'extreme_recovery_test' by @jcsp
• PR #14307 [v22.3.x] cloud_roles: redact fields from request string representation by @abhijat

Full Changelog: https://github.com/redpanda-data/redpanda/compare/v22.3.24...v22.3.25

Features

• Allow template prefix in advertised addresses by @RafalKorepta in #14204

Bug Fixes

• Fix a bug due to which the upload loop would terminate early if it failed to become in sync with the log
  at start-up. by @VladLazar in #14250
• Prevent compacted re-uploads for a given segment range from happening multiple times. by @VladLazar in #14182
• When serviceAccount.name is omitted in the Redpanda resource, we now fall back to the default in the Redpanda Helm chart. by @JakeSCahill in #14103
• When storage.persistentVolume.enabled is omitted in the Redpanda resource, we now fall back to the default in the Redpanda Helm chart. by @JakeSCahill in #14039
• #13973 Stop logging errors on shutdown of the tiered storage subsystem by @VladLazar in #13987
• #14028 Fix a quirk in our serialisation and deserialisation framework which could cause comparisons
  to ignore members of "wire" types. by @VladLazar in #14029
• #14120 Redpanda-Operator: multiple fixes to api by @alejandroEsc in #14121
• #14123 rpk topic describe-storage will now report the number of segments in the STM region of the cloud log, rather than 0. by @andrwng in #14124
• #14332 Fix a bug where the alter configs API would accept even-valued replication factors by @oleiman in #14334
• #14390 Fix a bug that could lead to crashes with double registration of group assertion after trying to create a lot of topics in a single CreateTopics request when some of these topics could not be allocated (e.g. due to bumping into partition count limits). by @ztlpn in #14392

Improvements

• #14060 Faster node decommissioning by @mmaslankaprv in #14061

• #14105 Kafka: Avoid very large contiguous allocations during fetch. by @BenPope in #14106

• PR #12159 [v23.2.x] ducktape: set small upload interval in retention_policy_test by @andrwng

• PR #13937 [v23.2.x] only run PR checks when needed by @joejulian

• PR #13978 [v23.2.x] Make maintenance mode opaque to the partition balancer by @rystsov

• PR #14001 [v23.2.x] tests/fixture: wait for controller quiescent state by @mmaslankaprv

• PR #14022 [v23.2.x] add github workflow for triggering promotion steps in bk by @nk-87

• PR #14037 [v23.2.x] archival: Audit meta in adjacent merger uploads by @Lazin

• PR #14045 [v23.2.x] ducktape: Exclude replaced segments when checking size by @abhijat

• PR #14055 [v23.2.x] Bump go version in k8s part of mono repo by @RafalKorepta

• PR #14082 [v23.2.x] Issue/13861/reduce api.github.com calls to fetch releases by @andijcr

• PR #14110 [v23.2.x] rptest: Add S3 internal server error to allowed log list by @Lazin

• PR #14113 [v23.2.x] cloud_storage/ducktape: Wait for chunk observer to stop by @abhijat

• PR #14159 [v23.2.x] rpk: bump go dependencies. by @r-vasquez

• PR #14173 [v23.2.x] operator: fix race condition in test by @joejulian

• PR #14185 [v23.2.x] s3_client: Check if response contains error by @Lazin

• PR #14190 [v23.2.x] ducktape/cloud_storage: Wait for topic info by @abhijat

• PR #14193 [v23.2.x] ducktape/cloud_storage: Wait for at least two spillovers by @abhijat

• PR #14199 [v23.2.x] Operator-Redpanda: remove omitempty from users list for sasl by @alejandroEsc

• PR #14209 [v23.2.x] scale tests: fix sizing variable name by @gousteris

• PR #14229 [v23.2.x] use the new controller factory shims for fluxcd by @joejulian

• PR #14272 [v23.2.x] operator: patch back in the checksum field into the flux crds by @joejulian

• PR #14278 [v23.2.x] operator: move kubebuilder comments back by @joejulian

• PR #14295 [v23.2.x] Fix for CI Failure (invalid timer interval) in UsageTest.test_usage_metrics_collection by @graphcareful

• PR #14296 [v23.2.x] operator: add readiness check to Cluster statefulset by @joejulian

• PR #14301 k8s: Fix go mod sum by @RafalKorepta

• PR #14308 [v23.2.x] cloud_roles: redact fields from request string representation by @abhijat

• PR #14310 [v23.2.x] k8s: Add secretRef to listener TLS definition by @RafalKorepta

• PR #14342 [v23.2.x] k8s: Regenerate Redpanda custom resource definition by @RafalKorepta

• PR #14343 [v23.2.x] k8s: remove users as required in auth.sasl by @alejandroEsc

• PR #14361 [v23.2.x] k8s: Bump grpc dependnecy by @RafalKorepta

• PR #14379 [v23.2.x] k8s: helm-v2 test fixed, and description for redpanda clusterspec updated by @alejandroEsc

Full Changelog: https://github.com/redpanda-data/redpanda/compare/v23.2.12...v23.2.13

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

Bug Fixes

• 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

Improvements

• #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: https://github.com/redpanda-data/redpanda/compare/v23.1.18...v23.1.19

Bug Fixes

• Allow console ingress reconciliation using Redpanda custom resource by @RafalKorepta in #13858
• 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 #13867
• #13492 Fixes an issue where autocreated topics would not reflect cluster-level cloud storage config by @oleiman in #13853
• #13813 External struct of the cluster spec was missing a few fields, notably SourceRanges and Annotations. by @alejandroEsc in #13814
• #13830 fixes a use_after_free due to interaction with iterators and suspension points. by @andijcr in #13893
• #13914 Fix overwriteTopicName usage. by @RafalKorepta in #13915
• #13954 Schema Registry: GET /subjects no longer returns a subject that has no versions or only deleted versions (unless ?deleted=true) by @BenPope in #13955
• do not allow unsafe decommission procedure if annotation was not changed by @RafalKorepta in #13901

Improvements

• Partition balancer now takes partition size reclaimable by space management into account. by @ztlpn in #13784
• #13670 Changes the log level for when rpk clients disconnect from the cluster to debug by @vuldin in #13671
• #13721 Export per handler produce/fetch metrics on the public_metrics endpoint by @StephanDollberg in #13731
• #13813 Enterprise struct with new license fields also created, deprecating the old license keys by @alejandroEsc in #13814
• #13926 Flushing Raft underlying logs will allow stms to create snapshots and minimize amount of data read during redpanda startup by @mmaslankaprv in #13928
• #13926 raft_election_timout_ms change does not require restart now by @mmaslankaprv in #13928
• #13960 Schema Registry: Improve error message for invalid compatibility level by @BenPope in #13961
• kafka/client logs now include the client identifier (e.g., schema_registry_client, pandaproxy_client) by @BenPope in #13969
• redpanda/kafka: Eliminate some copies in kafka responses (in particular, metadata) by @BenPope in #13881

Full Changelog: https://github.com/redpanda-data/redpanda/compare/v23.2.11...v23.2.12

Bug Fixes

• Fix race leading to assertion on materialisation of cloud storage segments by @VladLazar in #13823

• PR #13726 [v23.2.x] rpk: fix typo in decommission-status by @daisukebe

Full Changelog: https://github.com/redpanda-data/redpanda/compare/v23.2.10...v23.2.11

Features

• #13080 Schema Registry: Support GET /schemas/ids/{id}/subjects by @oleiman in #13082

Bug Fixes

• Redpanda will now allocate less contiguous memory when scanning a tiered storage cache directory that contains many empty directories. by @andrwng in #13097

• Schema Registry: Check compatibility iff the schema is a new version for the subject by @BenPope in #12963

• #11721 Memory consumption for housekeeping on compacted topics is reduced by @jcsp in #11722

• #12925 formatting a client/fetch_session would launch an exception, clobbering the output by @andijcr in #12928

• #12987 Schema Registry: GET /subjects/{subject}/versions/{version}/referencedby now returns a 404 if the subject or version doesn't exist by @BenPope in #12988

• #13540 rpk: prevent panic when using a directory instead of a filename in rpk redpanda tune --output-script by @r-vasquez in #13548

• fixed very rare error when consumer group recovery was incomplete by @mmaslankaprv in #13676

• PR #11911 [v22.3.x] Fixed querying for the node in decommission test by @mmaslankaprv

• PR #12186 v22.3.x rpk: bump franz-go deps by @twmb

• PR #12299 [v22.3.x] s/segment_set: recover base_index of last_clean_segment by @rystsov

• PR #12770 [v22.3.x] co_await mutex hold in conn_quota by @rockwotj

• PR #12875 [v22.3.x] 3rdparty: Update seastar for semaphore_units fix by @BenPope

• PR #13045 [v22.3.x] gha: bump quill to 0.4.1 by @andrewhsu

• PR #13136 [v22.3.x] rptest: improve timequery below start offset test by @andrwng

• PR #13170 [v22.3.x] k8s: go mod tidy by @andrewhsu

• PR #13315 [v22.3.x] Handle gate closing exceptions in replicate_entries_stm by @rockwotj

• PR #13408 [v22.3.x] gha: run go test on PRs to release branches by @andrewhsu

• PR #13690 [v22.3.x] deprecate operator helm chart by @joejulian

• PR #13692 rpk: fix lint issue in v22.3.x by @r-vasquez

• PR #13761 [v22.3.x] rpk tune fix checking for tune script file by @gousteris

• PR #13764 [v22.3.x] ducky/partition_balancer_test: fix full_node_test by @ztlpn

Full Changelog: https://github.com/redpanda-data/redpanda/compare/v22.3.23...v22.3.24

Features

• Allow to move pods into new nodes if needed, and allow down scaling Redpanda custom resource. by @alejandroEsc in #13624

Bug Fixes

• Correctly respond to changes to managed=true/false annotation. by @alejandroEsc in #13723

• Fix handling of slow down responses for HeadObject requests by @VladLazar in #13622

• Responds more expediently to deletion requests of a redpanda resource. by @alejandroEsc in #13723

• #13053 fixed rare but possible deadlock that might lead to Redpanda crash when moving partitions back and forth between nodes. by @mmaslankaprv in #13674

• #13640 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 #13694

• if a node is in maintenance mode it will not be used to serve fetch requests when follower fetching is enabled. by @mmaslankaprv in #13677

• PR #13643 [v23.2.x] c/pb_backend: schedule tick immediately if health report is available by @mmaslankaprv

• PR #13666 [v23.2.x] c/topics_frontend: create partition constraints on allocator shard by @mmaslankaprv

• PR #13688 [v23.2.x] deprecate operator helm chart by @joejulian

• PR #13746 [v23.2.x] Txn sessions thrashing: expire all old txs instead of one at a time by @rystsov

• PR #13762 Backport PRs 12888 and 12479 by @abhijat

Full Changelog: https://github.com/redpanda-data/redpanda/compare/v23.2.9...v23.2.10

Features

• #13380 rpk: add move-status to show ongoing partition movements by @daisukebe in #13381
• #13487 New Topic custom resource that can manage single topic by @RafalKorepta in #13488
• add support to the Console API (deprecated) for a user-specified ServiceAccount by @joejulian in #13357
• 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 #13533

Bug Fixes

• Allow Redpanda custom resource to configure console subchart by @RafalKorepta in #13606
• Fix issues with customers setting empty the list of users and using a predefined secret. by @alejandroEsc in #13530
• Fix migration of superuser list. by @alejandroEsc in #13576
• #13378 rpk/bugfix: enable back setting TLS struct values in set commands by @r-vasquez in #13382
• #13594 Removing spec.clusterSpec.monitoring.scrapeInterval as required. by @alejandroEsc in #13595
• rpk profile set now supports multiple comma separate values per key (key=value1,value2) by @twmb in #13375
• rpk: prevent panic when using a directory instead of a filename in rpk redpanda tune --output-script by @r-vasquez in #13539
• fixed rare situation in which follower recovery stuck as the follower state was incorrectly updated #13495

Improvements

• Show the true count of leaderless and under-replicated partitions in the admin API. by @travisdowns in #13398
• #13326 Tiered-storage partitions will now upload at least once an hour. See the cloud_storage_segment_max_upload_interval_sec cluster configuration for more details. by @andrwng in #13327
• #13433 The broker now logs errored Kafka API responses at the DEBUG level. by @NyaliaLui in #13434

Full Changelog: https://github.com/redpanda-data/redpanda/compare/v23.2.8...v23.2.9

Features

• adds usage metrics to each Kafka request handler by @ballard26 in #13344

Bug Fixes

• Fix OOM crashes for compacted topics and speed up filtering of compaction keys by up to 10x by @VladLazar in #13167

Improvements

• adds a kafka_handler_latency_seconds metric to the internal metrics endpoint that provides per-method-type latency histograms by @ballard26 in #13344

• PR #13128 [v23.1.x] Add thrashing of old txn sessions to prevent OOM by @rystsov

• PR #13201 [v23.1.x] c/members_manager: validate if node exists in the cluster before update by @mmaslankaprv

• PR #13311 [v23.1.x] Handle gate closing exceptions in replicate_entries_stm by @rockwotj

Full Changelog: https://github.com/redpanda-data/redpanda/compare/v23.1.17...v23.1.18

~
~
~

Features

• Repanda Tiered Storage is now compatible with Azure Storage accounts which have Hierarchical Namespaces enabled by @VladLazar in #13103
• #13081 Schema Registry: Support GET /schemas/ids/{id}/subjects by @oleiman in #13084
• add support to the Cluster API (deprecated) for a user-specified ServiceAccount by @joejulian in #13000

Bug Fixes

• Allow single reader scan the entire partition if spillover is enabled. by @Lazin in #13169
• Redpanda will now allocate less contiguous memory when scanning a tiered storage cache directory that contains many empty directories. by @andrwng in #13096
• #13162 Fix OOM crashes for compacted topics and speed up filtering of compaction keys by up to 10x by @VladLazar in #13163
• fix a bug in the operator's Cluster controller where scaling a cluster would cause the pods to be rolled by @joejulian in #13111
• fixed discrepancy between node list reported in /brokers and /cluster_config/status endpoints by @mmaslankaprv in #13104
• rpk will now search for a byoc plugin installed in the user's $PATH. by @r-vasquez in #13003

Improvements

• Timequeries (i.e ListOffsets requests) that land in the cloud log now use an index to speed up the search
  and reduce the number of hydrated bytes required to serve the query. On average, a time query will have to download 4 times less data (if using the default segment and chunk size) by @VladLazar in #13105
• #12884 rpk: show allocation failures in decommission-status by @daisukebe in #12885
• #13147 rpk: decommission-status supports -H option to show partition size in a human-readable format by @daisukebe in #13148
• PR #13063 [v23.2.x] cloud_storage: Force exhaustive trim when fast trim fails by @abhijat
• PR #13107 [v23.2.x] k8s: allow operator to create servicemonitors by @alejandroEsc
• PR #13113 [v23.2.x] k8s: update api with latest changes in helm chart by @alejandroEsc
• PR #13114 [v23.2.x] Improved segment appender write merging by @travisdowns
• PR #13123 [v23.2.x] Add thrashing of old txn sessions to prevent OOM by @rystsov
• PR #13127 Update seastar git SHA by @ballard26
• PR #13149 [v23.2.x] [storage]: make space management chill out by @dotnwat

Full Changelog: https://github.com/redpanda-data/redpanda/compare/v23.2.7...v23.2.8

Features

• #13083 Schema Registry: Support GET /schemas/ids/{id}/subjects by @oleiman in #13085

Bug Fixes

• Avoids a race between appends and application of segment.ms that could result in a crash. by @andrwng in #12715
• Redpanda will now allocate less contiguous memory when scanning a tiered storage cache directory that contains many empty directories. by @andrwng in #13098
• Schema Registry: Check compatibility iff the schema is a new version for the subject by @BenPope in #12962
• #12926 formatting a client/fetch_session would launch an exception, clobbering the output by @andijcr in #12929
• #12985 Schema Registry: GET /subjects/{subject}/versions/{version}/referencedby now returns a 404 if the subject or version doesn't exist by @BenPope in #12986
• fixed very rare error when consumer group recovery was incomplete by @mmaslankaprv in #12978

Improvements

• #10635 Automated topic recovery will be faster when operating on buckets with a large number of objects. by @andrwng in #10651
• reduced memory pressure and fragmentation by @mmaslankaprv in #12948
• PR #12436 [v23.1.x] txns: fix describe producers api to return correct last timestamp by @rystsov
• PR #12874 [v23.1.x] 3rdparty: Update seastar for semaphore_units fix by @BenPope
• PR #12914 [v23.1.x] storage: Make segment_appender::_flush_ops a fragmented_vector by @StephanDollberg
• PR https://github.com/redpanda-data/seastar/pull/61 : Fix a bug with the epoll reactor backend potentially causing cpu spinning at 100%

Full Changelog: https://github.com/redpanda-data/redpanda/compare/v23.1.16...v23.1.17

Bug Fixes

• Fix time query behaviour in situation when the timestamp doesn't match any segment by @Lazin in #12930
• Schema Registry: Check compatibility iff the schema is a new version for the subject by @BenPope in #12961
• #12674 fixes reporting maintenance mode status for nodes which are offline by @mmaslankaprv in #12675
• #12924 formatting a client/fetch_session would launch an exception, clobbering the output by @andijcr in #12927
• #12942 in some interleaving of appends + prefix_truncate + lower_bound, the wrong internal hint would be accessed by @andijcr in #12943
• #12974 Fix consistency violation caused by compacted segment reupload by @Lazin in #12973
• #12983 Schema Registry: GET /subjects/{subject}/versions/{version}/referencedby now returns a 404 if the subject or version doesn't exist by @BenPope in #12984
• fixed very rare error when consumer group recovery was incomplete by @mmaslankaprv in #12979

Improvements

• Cardinality of the set of metrics exposed by the tiered-storage decreased by @Lazin in #12993
• #12649 kafka: Abort fetch and list_offsets (timequery) when the client disconnects for timely resource cleanup. by @BenPope in #12981
• #12932 Make enforcement of cloud_storage_max_segment_readers and cloud_storage_max_materialized_segments_per_shard strict by @VladLazar in #12933
• #12968 rpk container improvements: rpk now will search for your docker context if no DOCKER_HOST is set when running rpk container commands by @r-vasquez in #12969
• faster convergence of partition balancer in face of topic operations by @mmaslankaprv in #12964
• reduced memory pressure and fragmentation by @mmaslankaprv in #12947

Full Changelog: https://github.com/redpanda-data/redpanda/compare/v23.2.6...v23.2.7

Bug Fixes

• Avoids a crash when attempting to create a read replica topic while cloud storage is not configured. by @andrwng in #11807
• Fix for OffsetForLeaderEpoch returning a value for when the requestedEpoch is larger then the highest known. by @graphcareful in #12122
• Fix for OffsetForLeaderEpoch returning the current leaders epoch (term) instead of the requested. by @graphcareful in #12122
• #10777 Fixes an issue with the archival upload path that could contribute to data loss when archival metadata updates took a long time to be replicated and applied. by @andrwng in #12690
• #11546 Fixed not being able to elect a leader in situation when only voter is in maintenance mode by @mmaslankaprv in #11667
• #11600 net: Fix a rare crash during shutdown of a failed connection with outstanding requests by @BenPope in #11601
• #12412 #12414 schema_registry: Strip redundant namespaces in Avro to improve schema lookup by @BenPope in #12419
• #12696 Schema Registry: Report 40408 if a subject config request fails by @BenPope in #12698
• fixes log segments being evicted too early by @mmaslankaprv in #12191
• PR #11308 [v22.3.x] Fix aborted txs being visible by @rystsov
• PR #11489 [v22.3.x] node_decommissioning_test: fix order of throttle and decommission by @ztlpn
• PR #11490 [v22.3.x] c/partition_balancer: reduce log level in scale tests by @ztlpn
• PR #11540 [v22.3.x] tests: fixed incorrectly used ^ operator in scale tests by @mmaslankaprv
• PR #11557 [v22.3.x] ducktape: improve pid dectection v2 by @rockwotj
• PR #11687 [v22.3.x] net: ss::nested_exception might be a disconnect_exception by @BenPope
• PR #11711 net/tests: Register a BOOST_TEST_MODULE by @BenPope
• PR #11762 [v22.3.x] rpc: Handle errors while holding a gate by @rockwotj
• PR #11804 [v22.3.x] add cdt script to install dt deps on cdt infra by @gousteris

Improvements

• Lower the TCP keepalive timeout to reap dead/idle connections faster and claim back resources by @StephanDollberg in #11775
• Redpanda will now be less verbose when logging certain operations in partitions with many local segments. by @andrwng in #12539
• #11646 Improved efficiency in encoding tag values in kafka wire protocol by @michael-redpanda in #11647
• #12050 Redpanda will now gracefully handle badly formatted SCRAM authentication messages by @michael-redpanda in #12060
• #12408 schema_registry: Improve the ordering of protobuf files by @BenPope in #12410
• #12855 Schema Registry: Introduce GET subjects/{subject}/versions/{version}/referencedby by @BenPope in #12856
• PR #12002 [v22.3.x] Add metric to monitor seq table size by @rystsov
• PR #12007 [v22.3.x] Fix txn consume group issues leading to undefined behavior by @rystsov
• PR #12114 acl: Avoid allocations when checking implied ops by @StephanDollberg
• PR #12692 [v22.3.x] c/create_topics: improvements to create_topic workflow #12531 by @bharathv

Full Changelog: https://github.com/redpanda-data/redpanda/compare/v22.3.22...v22.3.23

Bug Fixes

• Fixes a bug in usage based billing where timer doesn't fire at configured interval. by @graphcareful in #12835
• Fixes a bug in usage based billing where windows with inconstant begin/end timestamps are not tossed from the result set. by @graphcareful in #12835
• #12764 Fixes a bug in usage framework that where unhandled exceptions may be observed by @graphcareful in #12766
• PR #12769 [v23.2.x] co_await mutex hold in conn_quota by @rockwotj
• PR #12774 [v23.2.x] Await acquiring the manifest lock by @rockwotj
• PR #12790 [v23.2.x] Optimize send path error handling by @rockwotj

Improvements

• Redpanda will now be less verbose when logging certain operations in partitions with many local segments. by @andrwng in #12537
• #12853 Schema Registry: Introduce GET subjects/{subject}/versions/{version}/referencedby by @BenPope in #12854
• PR #12768 [v23.2.x] Context per reconciler by @joejulian

Full Changelog: https://github.com/redpanda-data/redpanda/compare/v23.2.5...v23.2.6

~

Bug Fixes

• Fixes a bug in usage based billing where timer doesn't fire at configured interval. by @graphcareful in #12836
• #10775 Fixes an issue with the archival upload path that could contribute to data loss when archival metadata updates took a long time to be replicated and applied. by @andrwng in #12438
• #12693 Schema Registry: Report 40408 if a subject config request fails by @BenPope in #12694
• #12765 Fixes a bug in usage framework that where unhandled exceptions may be observed by @graphcareful in #12767

Improvements

• #12851 Schema Registry: Introduce GET subjects/{subject}/versions/{version}/referencedby by @BenPope in #12852
• PR #12578 [v23.1.x] txns: fix list txns failure when it's used before find_coordinator by @rystsov
• PR #12597 [backport] [v23.1.x] c/create_topics: improvements to create_topic workflow #12506 by @bharathv
• PR #12771 [v23.1.x] co_await mutex hold in conn_quota by @rockwotj

Full Changelog: https://github.com/redpanda-data/redpanda/compare/v23.1.15...v23.1.16

Bug Fixes

• Avoids a race between appends and application of segment.ms that could result in a crash. by @andrwng in #12623
• Fixes a race between remote storage spillover that could result in fewer Kafka records being returned to consumers. by @andrwng in #12624
• #12695 Schema Registry: Report 40408 if a subject config request fails by @BenPope in #12697
• #12702 Fix race condition in tiered-storage which could cause a transient read error (missing offset) by @Lazin in #12703
• #12738 split pod annotation and label functions to prevent unexpected decommissioning of brokers by @joejulian in #12739

Improvements

• Add the ability to disable hints via a config flag by @andijcr in #12740
• PR #12172 [v23.2.x] Print output from Rpk commands that timeout in ducktape by @graphcareful
• PR #12388 [v23.2.x] tests: disable retention.ms in reader stress test by @jcsp
• PR #12512 [v23.2.x] dockerize lint clang-format by @gousteris
• PR #12527 [v23.2.x] rpk: fix flag name in help text of rpk generate app by @r-vasquez
• PR #12528 [v23.2.x] admin/observability/debugging: add stm information to partition debug endpoint by @bharathv
• PR #12556 [v23.2.x] tests: sort nodes before stopping them by @mmaslankaprv
• PR #12635 [v23.2.x] Fix segmentation fault in tx_gateway_frontend::expire_old_txs by @rystsov
• PR #12638 [v23.2.x] tests/cg-recovery: retry list groups by @NyaliaLui
• PR #12652 [v23.2.x] Do not compare fsid when checking if reports are equal by @mmaslankaprv
• PR #12736 k8s: use local crds when testing by @alejandroEsc
• PR #12742 [v23.2.x] test: check stderr for coordinator error messages by @dotnwat

Full Changelog: https://github.com/redpanda-data/redpanda/compare/v23.2.4...v23.2.5

Improvements

• Modifies usage metrics to be aligned by the configured interval by @graphcareful in #12593
• Re-enable separate fetch scheduling group by default. In the majority of cases as seen in cloudv2 it seems to be beneficial. by @StephanDollberg in #12591
• Redpanda will now be less verbose when logging certain operations in partitions with many local segments. by @andrwng in #12538
• PR #12298 [v23.1.x] s/segment_set: recover base_index of last_clean_segment by @rystsov
• PR #12523 [v23.1.x] transactions: fix ListTransactions to return open transactions by @rystsov

Full Changelog: https://github.com/redpanda-data/redpanda/compare/v23.1.14...v23.1.15

Features

• Report Redpanda Node ID will be reported as label and annotation in Pod metadata by @RafalKorepta in #12534

Bug Fixes

• Fixes a race between remote storage spillover that could result in fewer Kafka records being returned to consumers. by @andrwng in #12622
• #12517 Handle in Operator none existent Redpanda Node ID in Pod annotation by @RafalKorepta in #12518

Improvements

• Re-enable separate fetch scheduling group by default. In the majority of cases as seen in cloudv2 it seems to be beneficial. by @StephanDollberg in #12592
• #12525 Handle raft shutdown errors for delete record requests by @rockwotj in #12526
• #12561 rpk now prints the total lag of a group in rpk group describe and a new --print-lag-per-topic to aggregate the lag per topic. by @r-vasquez in #12562
• #12588 Modifies usage metrics to be aligned by the configured interval by @graphcareful in #12589
• #12618 Aggregate cloud_storage histograms by partition when the aggregate_metrics cluster property is set to true. by @ballard26 in #12620
• allow one attempt to upgrade of Redpanda resource and if fails allow to rollback in operator by @RafalKorepta in #12603