What's new in IBM Fault Analyzer for z/OS® V17.1

Latest service:

For information about the problems that have been addressed in each APAR, see Latest PTFs for IBM Fault Analyzer for z/OS.

This edition of the documentation provides information applicable to Fault Analyzer Version 17 Release 1. Changes to the latest release are indicated by a "|" change bar in the left margin of the page in the PDF format of the documentation.

October 2026 (V17R1M0, first edition)

This edition of the documentation contains minor clarifications and corrections, as well as the following significant changes relative to the previous version:

Modernizing the mainframe with Artificial Intelligence enhancements: In addition to traditional interactive and batch reanalysis, IBM Fault Analyzer for z/OS® 17.1.0 introduces two artificial intelligence capabilities to accelerate problem determination:
- IBM Fault Analyzer for z/OS Abend Intelligence: An automated, real-time, host-based analysis pipeline running natively on z/OS. When an application abnormally ends, IBM Fault Analyzer for z/OS Abend Intelligence automatically triggers an asynchronous background job (via the Notification user exit) to analyze the abend using a large language model (LLM), and output a Markdown-formatted reanalysis report. The pipeline supports both watsonx and Ollama as AI providers. For configuration details, see Installing and configuring IBM Fault Analyzer for z/OS Abend Intelligence.
- Asynchronous AI assistant integration using Model Context Protocol (MCP) server: IBM Fault Analyzer for z/OS® 17.1.0 supports integration with external AI assistants through an optional workstation-based Model Context Protocol (MCP) server. Running on Windows, the MCP server connects directly to the host using z/OSMF REST APIs, requiring no software installation on the mainframe LPAR. The server provides a set of tools that allow external AI assistants to locate history data sets, search history files, retrieve reanalysis reports, read SDSF job spool logs, and extract application source code.
  The MCP server also exposes a set of REST API services with a Swagger UI for direct scripting and manual testing if the server is run in HTTP mode. You can evaluate and test the server before connecting it to an AI client.
  For details, see Asynchronous AI analysis using the Model Context Protocol (MCP) server and Configuring the Model Context Protocol (MCP) Server.
- HCL Doc IQ, AI-powered search and assistance: With Version 17.1.0, we have integrated an AI-driven chat assistant built directly into the documentation portal. Powered by Agentic RAG and large language models (LLMs), HCL Doc IQ lets you ask technical questions in plain language and receive precise, context-aware answers instantly. The chat assistant eliminaties the need to manually search through topics and search results. All responses are generated exclusively from official product documentation to ensure maximum accuracy and reliability. Simply click the HCL Doc IQ icon in the corner of the portal to get started.
The new CompileDateCheck option specifies whether Fault Analyzer scans user programs in the event list to check if any were compiled within a specified number of days before the abend . If Fault Analyzer detects recently compiled user programs, the synopsis section of the fault analysis report includes a note. For details, see CompileDateCheck.
Two new commands have been added:
- HTTP command: displays formatted HTTP data located at a specified storage address. See HTTP for the complete command syntax and see Displaying formatted HTTP data (HTTP) for information about how to use this command and interact with the results.
- XML command: displays formatted XML data located at a specified storage address. See XML for the complete command syntax and see Displaying formatted XML (XML) for information about how to use this command and interact with the results.
Direct filtering of IMS programs when using IDIUTIL:
- The LISTHF, LISTHFDUP, and DELETE control statements of the IDIUTIL utility now support the IMS_PGM qualifier parameter. You can now list, filter, and delete fault entries by specifying an IMS program name.
  For more details, see LISTHF control statement, LISTHFDUP control statement, and DELETE control statement.
The ability to provide a user name or title when creating a fault entry from interactive MVS dump analysis has been provided. For details, see Creating a fault entry.