Downloads

The login command with force flag will no longer drop tokens from other profiles

PAUSED and PAUSING state added in decodable-cli output

Decodable pipelines expose their update_time.

ARM builds are now available for Linux

A bug in the decodable token refresh command has been fixed.

Declarative resource management: query and apply commands now support resource tags.

Declarative resource management now supports connectors that read or write from multiple streams.

The new token command, and its subcommands, allow you to retrieve an auth token from the CLI for scripting.

curl https://<accountName>.api.decoable.co/<endpoint> \
     -H "Authorization: Bearer $(decodable token access)"

Supported platforms:

Declarative resource management:

The --type and --flink-version flags are now required when creating custom pipelines.

Declarative resource management:

Declarative resource management:

De-duplicate Custom Pipeline file uploads

Declarative resource management: apply now supports custom pipelines.

Declarative resource management: query added to fetch and filter Decodable resources in apply-format YAML.

Declarative resource management: apply added to use YAML files to manage Decodable resources.

Secrets can now be made available to custom pipelines using the --secrets flag on pipeline create and pipeline update commands

Improved alignment of columns in list responses

Added the current state of pipelines and connections to the columns in list responses

Fixed a bug related to stream creation when an old schema is used

Removed file completions for CLI commands that don’t take a file argument

Tab completion is now supported for all commands. See decodable completion --help for instructions on how to set it up. Note: Dynamic completions of resource ids aren’t supported yet.

Added the decodable stream preview command for previewing data in a stream.

A warning will now be displayed if a deprecated CLI version is used, prompting to upgrade.

Fixed a bug concerning connection creation: When setting a sink stream via --stream-id flag, no fields need to be specified.

Fixed an error when creating connections with multi-column primary keys from the CLI.

Though there should be no user-facing impact, several commands were changed internally to account for the stricter control plane - data plane separation that was introduced alongside Bring Your Own Cloud.

Removed timestamps from error responses.

Fixed a bug where status codes weren’t reported correctly in error responses.

Enhanced schema management functionality has been added.

Responses now display - instead of <none> if no value is available.

Custom pipelines can now be updated using the pipeline update command. The --job-file and --job-arguments flags were added to this command (tech preview customers only)

Added custom pipeline properties to the pipeline get console output

Custom pipelines can now be created using the pipeline create command. To that end, the --job-file and --job-arguments flags were added (tech preview customers only)

The connection activate and pipeline activate commands now support setting the Task count and sizes via the --task-size flag.

Pipeline preview now works with change streams

You can configure the start position for pipelines and connections

Partition key support has been added to the stream create & update commands

The stream type field has been added to the decodable stream get output.

Fixed an issue that caused the login command to not store updated access tokens.

The activate command for pipelines & connections now support configurable parallelism via the --tasks flag.

Stream create & update commands now support user-defined watermarks for event-time processing!

Pipeline and connection get commands give increased visibility into your workloads by exposing throughput metrics and runtime error details.

Stream IDs are no longer required when creating connections. If no stream ID is given, a new stream is automatically created as the source/sink of the connection’s records.

Schema definition is no longer required when creating a connection if a stream ID is given. The connection schema will be implicitly defined to match the stream’s schema.

You no longer need to login again each time you switch profiles (or use the -p flag). Authentication and refresh tokens are now persisted for each profile individually.