What's new in 1.3.1.1

Cloud Native Link:

Cloud Native Link now supports deploying HCL Link on Kubernetes using a dedicated Helm chart. This cloud-native deployment model makes use of Kubernetes features such as container orchestration, self-healing, automated rollouts and rollbacks, service discovery, and persistent storage management. The new architecture organizes Link into Kubernetes services, pods, and volumes, enabling scalable, reliable, and consistent deployments across environments. Additionally, the default logging level in cloud-native environments has been changed from ALL to ERROR, and the connectors have been unbundled from the cloud-native images.

Adapters

Updates to existing adapter

Amazon S3 adapter: The Amazon S3 adapter now supports both permanent access keys and assume role–based authentication for accessing S3 objects. When assume role authentication is enabled, the adapter uses the provided ARN and duration to generate temporary credentials, which are automatically refreshed upon expiry. These temporary credentials are stored in a local persistence file, with an option to customize the storage directory. New command-line options (-AR, -ARN, -DU) allow users to configure assume role–based access for S3 Get and Put requests.
The adapter also supports the List Objects operation, allowing users to retrieve object listings directly from S3 buckets.
REST adapter: The REST Adapter now includes full multithreading support for parallel execution of REST requests, expanded response handling for all content types, and improved traceability with thread-level logging. In addition, HTTP connection pooling has been introduced, enabling connection reuse when the Connection: keep-alive header is present. These enhancements collectively improve performance, reduce latency, and provide better observability for high-volume REST integrations.
Azure Adapter: The Azure Adapter now supports the List Objects operation in Get actions, enabling retrieval of blob listings directly from Azure Storage. New advanced options include Key Prefix for filtering files by prefix or folder, and Recursive Listing for fetching objects from subfolders when needed. New command-line options (-KP, -R) simplify blob navigation and improve flexibility in Link flows.

Split Node Enhancements

The Split node supports dynamic fetch unit configuration, enabling the previously fixed batch size used for burst-mode requests to be defined at runtime through a flow variable. This enhancement provides greater flexibility for adjusting the fetch unit based on runtime conditions.

The map_batch_size setting is available when the data source is map.

If this setting is not configured, the fetch unit defined in the map is used.
If configured, the specified value is applied. This value can be an integer (for example, 5) or a flow variable (for example, %varMapBatchSize%).

Unit tests for this functionality are available in gt_bursting.cpp (Bursting.MapBatchSize*).

Ensure that the adapter used in the map supports the fetch unit functionality.

Core and Server Enhancements

This release includes fixes for thread-local variable issues across the Core and Server components. A conditional mechanism is also introduced to determine whether a generated map variable should function as a thread-local variable or as a flow variable.

REST Client Node

The REST Client node now supports multithreading when the input request mode is configured for multi-request, providing improved TPS (Transactions Per Second).

CDP and UI

A CDP update ensures that action names and connection names are now tenant-specific and no longer required to be unique across all tenants.