Highlights

The SnapLogic February 2025 Release includes the following highlights:

AgentCreator

• Tool Calling Support:

  • Added Tool Calling support to OpenAI for Assistants

  • Added Tool Calling support to Azure OpenAI for Assistants

  • Added Tool Calling support to Google GenAI for Gemini

• Vector Store Support:

  • Added support to manage files and vector stores for Azure OpenAI

  • Added support to create vector stores for OpenAI

• Added support to specify and configure an Azure AI Search Data source for Azure OpenAI Chat Completions usage

• Added support for removing records and indexes from Pinecone.

Snaps

• Introduced the Amazon Athena Snap Pack

• Introduced the gRPC Client to the API Suite Snap Pack

• Introduced the Oracle CDC Snap Pack

Upgrading to this Release

Security Update

Admin Manager

New Features

• In environments with Enhanced Email Encryption, Environment admins can now enable or disable the feature in Admin Manager. When enabled, individual users must configure their own account to receive encrypted emails. Learn more.

AgentCreator

New Features

• Added the following Snaps to the Azure OpenAI LLM Snap Pack:

  • Azure OpenAI Assistants Submit Tool Outputs: Submits a list of tool outputs to be fed back into the model as input for further processing by calling the Submit tool outputs to run the API.

  • Azure OpenAI Assistants Tool Call Router: Routes the input run object into two output views. The first output view passes the run object from the input, which also consists of an empty array, and the second output view includes the list of tools to be called.

  • Azure OpenAI Add Vector Store File: Adds an existing file from Azure OpenAI storage into the specified vector store using the vector store and file IDs.

  • Azure OpenAI Create Vector Store: Creates a new vector store associated with your account.

  • Azure OpenAI List Files: Retrieves and lists all files associated with the Azure OpenAI account.

  • Azure OpenAI List Vector Store Files: Retrieves and lists all vector store files based on the specified vector store ID from Azure OpenAI.

  • Azure OpenAI Remove Vector Store File: Removes a file from the specified vector store using the file ID from Azure OpenAI.

  • Azure Data Source AI Search: Enables search and retrieval across indexed datasets using Azure AI Search, allowing efficient query execution and intelligent information retrieval.

  • Azure Data Source API Key Account: Allows authentication using an API key to access Azure AI Search services, enabling secure and efficient search operations.

  • Azure User Assigned Managed Identity Account: Supports authentication via Azure Managed Identity, providing secure, identity-based access to Azure AI Search.

• Added the following Snaps to the Google GenAI LLM Snap Pack:

  • Google Gemini API Tool Calling: Provides external tools for the model to call, supplying internal data and information for the model's response.

  • Google GenAI Function Generator: Generates a tool definition in the output to inform the model about the tools that you want to call.

  • Google GenAI Function Result Generator: Formats the results generated by user-run functions into a unified structure.

• Added the following Snaps to the OpenAI LLM Snap Pack

  • OpenAI Create Vector Store Snap to create a new vector store associated with your account.

  • OpenAI Assistants Submit Tool Outputs: Submits a list of tool outputs to be fed back into the model as input for further processing by calling the Submit tool outputs to run API.

  • OpenAI Assistants Tool Call Router: Routes the input run object into two output views. The first output view passes the run object from the input, which also consists of an empty array. The second output view includes the list of tools to be called.

• Added the following Snaps to the Pinecone Snap Pack:

  • Pinecone Delete Index: Deletes indexes in a user's Pinecone service.

  • Pinecone Delete Records: Deletes records in a Pinecone index.

Enhancements

• Amazon Bedrock LLM:

  • Enhanced the Amazon Bedrock Function Result Generator Snap with the Metadata field set for adding key-value pairs to function results, enabling seamless downstream processing.

  • The Amazon Bedrock Function Result Generator Snap now supports boolean values from the upstream Snap in the Is Error field, which previously supported only pipeline parameters. If the field is left empty, the Snap issues a lint warning to indicate that the default value (false) is used.

  • The SdkNumber data type returned by the Amazon Converse API tool is now converted to BigDecimal data type which is supported by the SnapLogic Platform, to ensure seamless processing of numeric inputs.

• Azure OpenAI LLM:

  • Enhanced the Create and Run Thread and Run Thread Snaps to support the tool calling capability specifically for the assistants in the Azure OpenAI Snap Pack.

• Amazon Bedrock LLM, Azure OpenAI LLM, and OpenAI LLM:

  • Enhanced the Amazon Bedrock Function Generator, Azure OpenAI Function Generator, and OpenAI Function Generator Snaps to validate the names of functions based on the individual providers' requirements.

  • Enhanced the Azure OpenAI Function Result Generator and OpenAI Function Result Generator Snaps with the Metadata field set that supports custom key-value pairs configuration. This feature enables you to pass specific metadata from the function pipeline, such as function execution details captured in logs.

• OpenAI LLM: Enhanced the Create and Run Thread and Run Thread Snaps to enable the assistant's tool calling capability.

• Azure OpenAI LLM: Enhanced the following accounts in the Snap Pack with the Read timeout (seconds) field that enables the Snap to wait for a response from Azure:

  • Azure OpenAI API Key Account

  • Azure OpenAI Entra ID Application Account

  • Azure OpenAI Entra ID User Account

  • Azure OpenAI Entra Token Account

• Enhanced the Azure OpenAI Chat Completions Snap to support Azure AI Search Data Sources, allowing data source configurations and dependency management to optimize search and response processing.

API Management

Fixed Issues

• Logging into the Developer Portal with a username that starts with a leading hash character ( # ) no longer results in a 404 user not found error.

AutoSync

Enhancements

• The PostgreSQL and Databricks target endpoints now support writing error records to a table.

• In environments that use Enhanced Account Encryption, AutoSync data pipelines must run on a Groundplex. For new environments that subscribe to AutoSync, an AutoSync Cloudplex is no longer provisioned. In environments with an existing subscription, the AutoSync Cloudplex will no longer be available to select.

Fixed Issues

• The Azure Synapse source endpoint now supports the Incremental and SCD2 load types by allowing you to choose the field that tracks the last modified timestamp.

Documentation Enhancements

Migrated Help Articles

The following help articles have recently been migrated to https://docs.snaplogic.com from the legacy documentation site:

• https://docs.snaplogic.com/snaps/snaps-machine-learning/sp-ml-analytics/sp-ml-analytics-about.html

• https://docs.snaplogic.com/snaps/snaps-machine-learning/sp-ml-natural-language-processing/sp-ml-natural-language-processing-about.html

• https://docs.snaplogic.com/snaps/snaps-core/sp-snaplogic-metadata/sp-snaplogic-metadata-about.html

• https://docs.snaplogic.com/snaps/snaps-catalog.html

Monitor

Enhancements

• The documentation for node metric graphs now includes the time granularity for downsampling data. The granularity depends on the time period that you select for the graphs.

Fixed Issues

• In the Asset catalog, the View executions link on the task details pane now opens the Pipeline executions page with the correct executions for that task.

• The Pipeline executions page now shows the correct results when opened from the Asset catalog. If you click the View executions link from a pipeline’s detail page in the Asset catalog, only executions for the selected pipeline display.

Known Issues

• Multiple manual refreshes of the Pipeline executions page in a short time period can cause a display error. Workaround: Wait a few minutes and try again.

Platform

Enhancements

• Environment admins can control whether tasks expose a Cloud URL for Triggered Tasks. The project options dialog in Classic Manager includes a new Enable cloud URL checkbox, which applies to all tasks associated with pipelines in that project.

• The Git Integration settings for http://GitLab.com and self-managed GitLab now include a Namespaces field to specify the specific repositories for tracking SnapLogic assets. Limiting access to the appropriate repositories increases performance. The field is not required to ensure backward compatibility. Before using this feature, we recommend that Environment admins understand the correct way to reference GitLab groups, subgroups, and repositories.

• Two unused fields have been removed from the Git Integration self-managed GitLab settings: Http proxy and Https proxy.

Known Issues

• Auto-scaling in Kubernetes causes running pipelines to fail after a node is placed in maintenance mode.

Snaps

Upcoming

New

• Introduced the Amazon Athena Snap Pack with the following Snaps and Account:

  • Athena Bulk Load: Performs a bulk load operation from an external location to the target table.

  • Athena Query: Executes SQL queries on large datasets.

  • Athena Select: Executes a SELECT query on the Amazon Athena table.

  • Amazon Athena Account: Establishes a connection between Amazon Athena Snaps and Athena services.

• API Suite:

  • Added the gRPC Client Snap that enables you to interact with a gRPC (Google Remote Procedure Call) server to invoke its services. Learn more.

• Added the Oracle CDC Snap Pack with the following Snap and Account:

  • Oracle CDC Snap: Retrieves data from Oracle Database transaction logs. It performs Insert, Update, and Delete statements on a specific table within a specified date and time range.

  • Oracle CDC Account: Establishes a connection between the Oracle CDC Snap and Oracle CDC services.

Enhancements

• Amazon S3

  • Enhanced the S3 Account to access AWS S3 successfully in the ECS environment with IAM roles. Learn more.

• Google BigQuery:

  • The BigQuery Bulk Load (Cloud Storage) Snap now displays a lint warning when the bulk load process completes successfully but the temporary file deletion fails.

  • Enhanced the following Google BigQuery Snaps with implicit retry functionality to improve the reliability of CRUD operations by retrying after errors:

    • BigQuery Write

    • BigQuery Table Delete

    • BigQuery Table Data List

    • BigQuery Table Create

    • BigQuery Execute

    • BigQuery Dataset List

    • BigQuery Dataset Delete

    • BigQuery Dataset Create

    • BigQuery Table List

    • BigQuery Bulk Load (Cloud Storage)

• Databricks:

  • Enhanced the Databricks Account to support M2M authentication, allowing secure connections using a Client ID and Client secret.

• Enhanced the following Databricks Snaps to support the Unity catalog that enables you to access catalog data from the database.

  • Databricks - Bulk Load

  • Databricks - Delete

  • Databricks - Insert

  • Databricks - Merge Into

  • Databricks - Select

  • Databricks - Unload

• JDBC

  • Upgraded the JOOQ library of the JDBC Snap Pack from v3.9.1 to v3.17.x. Pipelines using the Generic - JDBC Snaps might be impacted by the jOOQ upgrade because the latest jOOQ version is not fully backward-compatible. Learn more.

• SAP

  • Enhanced the SAP Execute Snap with the Round to SAP decimal checkbox that enables the Snap to round the decimal value to match the number of decimal places defined in the SAP target field. This property only applies to the SAP packed decimal fields, which store data in the BCD (Binary Coded Decimal) format.

  • Enhanced the SAP iDoc Document Listener and SAP iDoc Listener Snaps for better error handling by ensuring that the system doesn't halt on connection errors, but does halt on on-connection errors. Now, critical non-connection errors will cause the system to halt, ensuring critical issues are addressed promptly.

• Teradata: Enhanced the following Teradata Snaps with the Query band field that logs the SnapLogic pipeline, project, and Teradata database information in Teradata to help associate metadata with each query when it is executed:

  • TPT FastExport

  • Teradata FastExport

  • Teradata FastLoad

  • Teradata Select

  • Teradata Execute

  • TPT Load

• The Snap trial option See add-on Snap Pack (in the Snap Catalog) is renamed to View Premium Snap Packs.

  • Watch how to request trials for Premium Snap Packs:

    Subscribe to the Premium Snap Packs.mp4

Behavior change

• SAP: Updated the behavior of the Halt on SAP error checkbox in the SAP iDoc Document Listener and SAP iDoc Listener Snaps for better error handling by stopping significant errors while allowing retries on connection errors. Instead of halting on any error, the system will now stop only on the first non-connection error. This change enhances the system's robustness and ensures critical issues are addressed promptly.

• Generic JDBC: Earlier, the ROWID columns were displayed in binary (Base64 encoded data) form in the output. With the 439patches29008, the Generic JDBC Execute and Select Snaps that integrate with the Oracle endpoint display the ROWID columns in string form in the output, which can be used for other operations downstream.

Known Issues

• Amazon Athena Snap Pack:

  • Athena Bulk Load:

    • Choosing the Overwrite option for the If table exists field can delete the Athena table and its associated S3 objects when the Snap execution starts. Use this option cautiously to prevent data loss.

  • Athena Select:

    • ARRAY, STRUCT, and MAP are not supported in SELECT queries. If used, they can result in incorrect values in the output documents.

    • This Snap can fail with a connection error if the account is configured with an IAM role, and, Role ARN.

    • This Snap uses the Athena JDBC driver to retrieve data. When using a Cross-account IAM role, the JDBC connection expires based on the role's duration. Large datasets requiring more time than this duration are not supported.

    • CSE_KMS encryption is not supported because of an issue in the Athena JDBC driver.

      Workaround: Use the Athena Query Snap with a SELECT query.

    • The second output view of this Snap cannot connect directly to the second input view of Athena Bulk Load Snap because Athena does not provide complete table metadata for all data types.

    • Query execution can fail on a table created with ORC files in LZO compression because of an Athena service defect.

• Athena Query

  • ARRAY, STRUCT, and MAP are not supported in SELECT queries. If used, it can result in incorrect values in the output documents.

  • Query execution can fail on a table created with ORC files in LZO compression because of an Athena service defect.

• When the Teradata FastLoad Snap pauses, both the target and error tables get locked.
  Workaround: You must manually drop these tables before retrying the fast-load job.

Deprecated

• The ForEach and TaskExecute Snaps are deprecated. If your existing pipelines use these Snaps, we recommend that you replace them with the Pipeline Execute Snap.

SnapGPT

Enhancements

• It’s now easier to select an expression from the list that SnapGPT generates. Each suggestion displays in a separate box: