For installation instructions check out the getting started guide.
Added
- New
mysql_cdc
input supporting change data capture (CDC) from MySQL. (@rockwotj, @le-vlad) - Field
instance_id
added tokafka
,kafka_franz
,ockam_kafka
,redpanda
,redpanda_common
, andredpanda_migrator
inputs. (@rockwotj) - Fields
rebalance_timeout
,session_timeout
andheartbeat_interval
added to thekafka_franz
,redpanda
,redpanda_common
,redpanda_migrator
andockam_kafka
inputs. (@rockwotj) - Field
avro.preserve_logical_types
for processorschema_registry_decode
was added to preserve logical types instead of decoding them as their primitive representation. (@rockwotj) - Processor
schema_registry_decode
now adds metadataschema_id
for the schema's ID in the schema registry. (@rockwotj) - Field
schema_evolution.processors
added tosnowpipe_streaming
to support side effects or enrichment during schema evolution. (@rockwotj) - Field
unchanged_toast_value
added topostgres_cdc
to control the value substituted for unchanged toast values when a table does not have full replica identity. (@rockwotj)
Fixed
- Fix a snapshot stream consistency issue with
postgres_cdc
where data could be missed if writes were happening during the snapshot phase. (@rockwotj) - Fix an issue where
@table
metadata was quoted for the snapshot phase inpostgres_cdc
. (@rockwotj)
Changed
- Field
avro_raw_json
was deprecated in favor ofavro.raw_unions
for processorschema_registry_decode
. (@rockwotj) - The
snowpipe_streaming
output now has better error handling for authentication failures when uploading to cloud storage. (@rockwotj) - Field
schema_evolution.new_column_type_mapping
forsnowpipe_streaming
is deprecated and can be replaced withschema_evolution.processors
. (@rockwotj) - Increased the default values for
max_message_bytes
andbroker_write_max_bytes
by using IEC units instead of SI units. This better matches defaults in Redpanda and Kafka. (@rockwotj) - Dropped support for postgres 10 and 11 in
postgres_cdc
. (@rockwotj)
The full change log can be found here.