Managing and understanding mainframe jobs can be challenging for students and new developers, especially when working in limited-access learning environments like IBM Z Xplore. In this session, we present Z-Agent, a lightweight system that integrates Zowe CLI with a Python backend to collect, analyze, and visualize job execution data from z/OS — all without requiring privileged SMF access.
Using only student-level tools, Z-Agent demonstrates how to extract valuable insights from job metadata, JCL outputs, and dataset attributes. We’ll show how to transform this raw data into dashboards that track job success rates, ABEND trends, and runtime analytics. The talk also highlights how AI-assisted log analysis (via Watson or OpenAI APIs) can help interpret job failures in natural language.
This project bridges modern DevOps principles with mainframe education, providing a practical example of how APIs and open-source tools can enhance mainframe learning and observability in hybrid environments.