v4.51.0

For installation instructions check out the getting started guide.

Added

• Field private_key added to ssh input and output to let users directly specify their private key contents in their config instead of writing it to a file (@ooesili)
• Field history added to ollama_chat processor to allow for chat history. (@rockwotj)
• Field history added to openai_chat_completion processor to allow for chat history. (@rockwotj)
• Field handle_logical_types added to parquet_decode input to provide better handling of Parquet logical types (@ooesili)
• New gateway input. (@Jeffail)
• New git input. (@weeco, @rockwotj)
• New text_chunker processor for splitting text for creating document vector embeddings. (@rockwotj)
• New aggregate operation added to the mongodb processor to provide support for aggregation pipelines. (@brknstrngz, @mihaitodor)
• New slack input reading from slack using socketmode. (@rockwotj)
• Option headers added to field type on the amqp_0_9 output. (@brknstrngz)

Fixed

• The azure_blob_storage input now drops targets_input notifications and emits a warning log message for blobs which have been deleted before Connect was able to read them. (@mihaitodor)

Changed

• Field type on the amqp_0_9 output now only enforces dots in routing keys and message types for topic exchanges. (@brknstrngz)

The full change log can be found here.

v4.51.0-rc8

For installation instructions check out the getting started guide.

NOTE: This is a release candidate, you can download a binary from this page.

Added

• Field private_key added to ssh input and output to let users directly specify their private key contents in their config instead of writing it to a file (@ooesili)
• Field history added to ollama_chat processor to allow for chat history. (@rockwotj)
• Field history added to openai_chat_completion processor to allow for chat history. (@rockwotj)
• Field handle_logical_types added to parquet_decode input to provide better handling of Parquet logical types (@ooesili)
• New gateway input. (@Jeffail)
• New git input. (@weeco, @rockwotj)
• New text_chunker processor for splitting text for creating document vector embeddings. (@rockwotj)
• New aggregate operation added to the mongodb processor to provide support for aggregation pipelines. (@mihaitodor)
• New slack input reading from slack using socketmode. (@rockwotj)

Fixed

• The azure_blob_storage input now drops targets_input notifications and emits a warning log message for blobs which have been deleted before Connect was able to read them. (@mihaitodor)

The full change log can be found here.

v4.51.0-rc7

For installation instructions check out the getting started guide.

NOTE: This is a release candidate, you can download a binary from this page.

Added

• Field private_key added to ssh input and output to let users directly specify their private key contents in their config instead of writing it to a file (@ooesili)
• Field handle_logical_types added to parquet_decode input to provide better handling of Parquet logical types (@ooesili)
• New gateway input. (@Jeffail)
• New text_chunker processor for splitting text for creating document vector embeddings. (@rockwotj)

Fixed

• The azure_blob_storage input now drops targets_input notifications and emits a warning log message for blobs which have been deleted before Connect was able to read them. (@mihaitodor)

The full change log can be found here.

v4.51.0-rc6

For installation instructions check out the getting started guide.

NOTE: This is a release candidate, you can download a binary from this page.

Added

• Field private_key added to ssh input and output to let users directly specify their private key contents in their config instead of writing it to a file (@ooesili)
• Field handle_logical_types added to parquet_decode input to provide better handling of Parquet logical types (@ooesili)
• New gateway input. (@Jeffail)
• New text_chunker processor for splitting text for creating document vector embeddings. (@rockwotj)

Fixed

• The azure_blob_storage input now drops targets_input notifications and emits a warning log message for blobs which have been deleted before Connect was able to read them. (@mihaitodor)

The full change log can be found here.

v4.51.0-rc5

For installation instructions check out the getting started guide.

NOTE: This is a release candidate, you can download a binary from this page.

Added

• Field private_key added to ssh input and output to let users directly specify their private key contents in their config instead of writing it to a file (@ooesili)
• Field handle_logical_types added to parquet_decode input to provide better handling of Parquet logical types (@ooesili)
• New gateway input. (@Jeffail)

The full change log can be found here.

v4.51.0-rc4

For installation instructions check out the getting started guide.

NOTE: This is a release candidate, you can download a binary from this page.

Added

• Field private_key added to ssh input and output to let users directly specify their private key contents in their config instead of writing it to a file (@ooesili)
• Field handle_logical_types added to parquet_decode input to provide better handling of Parquet logical types (@ooesili)
• New gateway input. (@Jeffail)

The full change log can be found here.

v4.51.0-rc3

For installation instructions check out the getting started guide.

NOTE: This is a release candidate, you can download a binary from this page.

Added

• Field private_key added to ssh input and output to let users directly specify their private key contents in their config instead of writing it to a file (@ooesili)
• Field handle_logical_types added to parquet_decode input to provide better handling of Parquet logical types (@ooesili)
• New rpingress input. (@Jeffail)

The full change log can be found here.

v4.51.0-rc1

For installation instructions check out the getting started guide.

NOTE: This is a release candidate, you can download a binary from this page.

Added

• Field private_key added to ssh input and output to let users directly specify their private key contents in their config instead of writing it to a file (@ooesili)
• Field handle_logical_types added to parquet_decode input to provide better handling of Parquet logical types (@ooesili)
• New rpingress input. (@Jeffail)

The full change log can be found here.

v4.50.0

For installation instructions check out the getting started guide.

Added

• Processor openai_chat_completion can now call tools that are defined as a series of additional processors. (@rockwotj)
• New bloblang function unicode_segments to split text based on unicode graphemes, words or sentences. (@rockwotj)

Fixed

• Output snowflake_streaming can now write float columns with NaN, -inf and inf values. (@rockwotj)

The full change log can be found here.

v4.49.1

For installation instructions check out the getting started guide.

Added

• Output snowflake_streaming has two new stats snowflake_register_latency_ns and snowflake_commit_latency_ns. (@rockwotj)

Changed

• Field snapshot_memory_safety_factor is now removed for input postgres_cdc, the batch size must be explicitly defined, the batch size default is 1000. (@rockwotj)
• Input postgres_cdc now supports intra-table snapshot read parallelism in addition to inter-table parallelism. (@rockwotj)

The full change log can be found here.