NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
Go modules · #908 by repository stars
Last release 2 days ago
07 Oct 2026
Ships on a steady schedule
a new release about every 1 weeks
Nearly every release is documented
notes for 42 of 42 stable releases
Nothing withdrawn
no release was ever pulled
2 years old
271 releases · first in 2024
migrated MSSQL driver to official microsoft/go-mssqldb
Full Changelog: v1.21.1...v1.21.2
aws_kinesis input with enhanced_fan_out no longer accumulates duplicate shard consumers when the pipeline applies sustained backpressure, which previously grew memory usage until the process was killed @matus-tomleinOne column per month.
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
kafka_franz output field broker_write_max_bytes misconfigured (introduced in version 1.21.0) @jem-davies
kafka_franz output field broker_write_max_bytes misconfigured (introduced in version 1.21.0) @jem-daviesFull Changelog: v1.21.0...v1.21.1
Optional Keys method for caches enabling consumers of cache resources to enumerate keys, caches that do not support key listing yield ErrKeyListingNot
Keys method for caches enabling consumers of cache resources to enumerate keys, caches that do not support key listing yield ErrKeyListingNotSupported @ecordellmemory, file, aws_s3 and gcp_cloud_storage caches support Keys @ecordellbucket field on aws_s3 output supports bloblang interpolation @jem-daviesexists which checks caches for keys @lublakextract_tracing_map and new_root_span_with_link on the gcp_pubsub input & inject_tracing_map on the gcp_pubsub output @jesperanzasolotransaction_isolation_level field to kafka & kafka_franz components @mbuckley2000send_ack field on nats input enables/disables sending acknowledgement back to nats' input subject reply address @ahart97memory, file, aws_s3 & gcp_cloud_storagebroker_read_max_bytes to kafka_franz input and broker_write_max_bytes & max_buffered_bytes options to kafka_franz output @jem-daviesparse_big_decimal rejects scales above 16383 to avoid unbounded big-integer work @aratz-lasafsevent preserves file create events when watching new subdirs @mattfaltynaws_s3_stream handles files smaller than S3's minimum multipart size @triddellgithub.com/hamba/avro/v2 with github.com/iskorotkov/avro/v2, temporarily disabling go install github.com/warpstreamlabs/bento/cmd/bento@<version> due to replace directive usage.Full Changelog: v1.20.0...v1.21.0
Keys method for caches enabling consumers of cache resources to enumerate keys, caches that do not support key listing yield ErrKeyListingNotSupported @ecordellmemory, file, aws_s3 and gcp_cloud_storage caches support Keys @ecordellbucket field on aws_s3 output supports bloblang interpolation @jem-daviesexists which checks caches for keys @lublakextract_tracing_map and new_root_span_with_link on the gcp_pubsub input & inject_tracing_map on the gcp_pubsub output @jesperanzasolotransaction_isolation_level field to kafka & kafka_franz components @mbuckley2000send_ack field on nats input enables/disables sending acknowledgement back to nats' input subject reply address @ahart97memory, file, aws_s3 & gcp_cloud_storagebroker_read_max_bytes to kafka_franz input and broker_write_max_bytes & max_buffered_bytes options to kafka_franz output @jem-daviesparse_big_decimal rejects scales above 16383 to avoid unbounded big-integer work @aratz-lasafsevent preserves file create events when watching new subdirs @mattfaltynaws_s3_stream handles files smaller than S3's minimum multipart size @triddellapache/pulsar-client-go updated to v0.21.0, picking up fixes for a consumer busy-spin deadlock, a connection panic on close, and a race that dropped messages before handler registration @slachiewiczgithub.com/hamba/avro/v2 with github.com/iskorotkov/avro/v2, temporarily disabling go install github.com/warpstreamlabs/bento/cmd/bento@<version> due to replace directive usage.Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
parse_big_decimal bloblang method for Kafka Connect / Debezium decimal decoding @aratz-lasa
parse_big_decimal bloblang method for Kafka Connect / Debezium decimal decoding @aratz-lasaaws_bedrock_invoke processor for invoking Amazon Bedrock models @louisdvs & @gregfurmankafka_franz components being routed to a single partition @jem-daviesbatchErr mutex locking in the elasticsearch_v2 output @jem-daviesFull Changelog: v1.19.0...v1.20.0
Nothing published for this version
Nothing published for this version
grpc_client input component @jem-davies
grpc_client input component @jem-daviessync_response added to nats input NATS @henrikschristensenendpoint config option added to gcp_bigquery output @Shareed2kparquet default encoding @georgeTobiaelasticsearch processor @Omar1H1json_number_type config option added to aws_dynamodb output enables JSON numbers to be either stored as dynamoDB Strings / Numbers @cthornernegotiate config options added to http components to enable SPNEGO auth @lublaktimestamp field added to gcp_bigtable enabling setting timestamps @gregfurmanrestart_backoff fields added to read_until input enable setting a backoff policy for when the child input restarts @aramperesaws_dynamodb output no longer drops successfully-built messages when another message in the batch fails request construction @cthorneraws_kinesis assign msgChan after succesful connection @richardartoulaws_kinesis input fixed a nil pointer panic when ListShards returned an error @richardartoulaws_kinesis input fixed ListShards pagination error @gregfurmanomit_if_empty config option added to aws_dynamodb enabling the correction of previously documented behaviour that missing paths are omitted rather than written as null @cthorneraws_kinesis input fixed shard consumer busy-loop and excess timer allocations noticible at high shard counts @richardartoulschema_registry_decode & schema_registry_encode fixed for schema's whose subject contained a / @PitchBlack07aws_kinesis input simplified EFO architecture to align with Amazon's Kinesis Client @matus-tomleinFull Changelog: v1.18.1...v1.19.0
avro scanner call Scan() before Read() @istairbn
avro scanner call Scan() before Read() @istairbnprotobuf processor buf schema registry poller goroutine leak @jem-daviesotlp dependency @jem-daviesFull Changelog: v1.18.0...v1.18.1
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
removed reference to deprecated 'Optional' field in cuegen @aebrahim
template processor enabling use of Go template syntax for message transformations @lublakfsevent input creates messages for file-system events with metadata pertaining to event @henrikschristensenduckdb_append output inserts rows into a DuckDB database using the Appender API @iamramtindigest_auth fields on http components enabling digest authentication @lublakoauth2 fields on kafka & kafka_franz enabling oauth2 authentication @vsl86 & @jem-daviesemit_unpopulated field to protobuf processor enabling emitting unpopulated fields with their default JSON values @matta-devchecksum_algorithm field added to aws_s3 when set Bento computes the checksum for s3 to check @seanbarzilayfile processor delete operation fixed for windows @henrikschristensensql_raw output @adrianhajkafka_franz input to attach logger regardless of consumer group being set @gregfurmankafka component library) @gitphillkafka_franz franz-go logs to surface at ERROR/WARN/INFO/DEBUG instead of ERROR/WARN/DEBUG/TRACE @gregfurmanFull Changelog: v1.17.0...v1.18.0
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
sanitize_namespace_names field to schema_registry_decode processor @jem-davies
sanitize_namespace_names field to schema_registry_decode processor @jem-daviesapi_key to elasticsearch_v2 output adding the option to use a elasticsearch API key to connect @alexthemayersextract_tracing_map and new_root_span_with_link fields to the kafka_franz input @danielavelez12kafka_timestamp_ms metadata fields to both kafka_franz and kafka @dnson & @sonnydinhgcfile processor @henrikschristensenFull Changelog: v1.16.2...v1.17.0
aws_kinesis input busy loop when batching.period is set @jem-davies
aws_kinesis input busy loop when batching.period is set @jem-daviesFull Changelog: v1.16.1...v1.16.2
add handling for authentication errors in SQL output components @adrianhaj
Full Changelog: v1.16.0...v1.16.1
datadog_logs output component @gregfurmanNothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Your coding agent can read these notes before it upgrades. Set up the MCP server →