PlatformManager 1.97.0

Version

1.97.0

Release Date

New Features and Improvements

Improvement Description


The file preview process now detects UTF-8 compatibility and displays a warning message if a non-UTF-8 encoded file is identified.

PFIM-10135

Public API endpoints are now accessible via the Model Context Protocol (MCP) to support integration with LLM tools.

PFIM-10114

The Failed Events tab now supports multi-tenant Integration Manager instances by displaying and filtering failed events by their specific Pricefx connection.

PFIM-10100

Engineers with eng-support accounts can now authenticate to partitions via Studio using Azure login.

PFIM-10081

The Payload section is now displayed only for action types that support it, specifically Logic and Route, to prevent user confusion regarding unsupported data delivery.

PFIM-10009

Partition administrators can now create a "Pricefx Client OAuth" connection on the Connections page of an integration in Platform Manager. An Authorize button obtains the token from the partition, so Integration Manager flows authenticate with OAuth instead of basic authentication. The client ID and the Platform Manager callback URL must already be registered on the partition.

PFIM-9931

The new GenAI Services page in Overview lists the AgentGen, MCP Server, and Gatekeeper instances with their status, environment, Kubernetes cluster, and address. Each service reports its status by heartbeat, and shows as Unknown after three hours without one. The row menu opens the Grafana logs, metrics, and traces for the selected service. Access requires the GenAI Services List read permission.

PFIM-9280

The Log History search filter now supports seconds in the timestamp selector, for more precise filtering.

PFIM-9157

The Show Definition screen for an event orchestration workflow now has Diagram and JSON tabs. Both show all listeners, filters, and actions without opening each listener.

PFIM-7658

Fixed Issues

Bug Description


The system fails to clone workflow orchestrations and scheduler jobs between partitions, resulting in a generic error message during the copy process.

PFIM-10211

The Instances list shows the Provider as BAREMETAL for clusters that run on AWS, and generates malformed monitoring links. The provider value is taken from stale or incomplete node records.

PFIM-10177

The Instances list returns a 500 error when a decommissioned cluster contains stale node registration records that lack associated asset data.

PFIM-10132

The system sends an invalid filter criterion during periodic polling for partition-job notification rules, causing a server-side ClassCastException and preventing alerts from functioning when a specific target is associated with an "ALL" tracker type.

PFIM-10128

The automated data upload progress bar stalls at 99% despite the file upload completing successfully.

PFIM-10110

The exported Access Report includes deactivated or unauthorized users who are not visible in the user interface.

PFIM-10105

The IM Upgrade Assistance tool incorrectly allows users to initiate unsupported direct updates to version 8.0.0 for instances running on Argo, causing the system to remain stuck in an upgrading state.

PFIM-10104