Rocket Delivers More DevOps Capabilities For IBM i
November 17, 2025 Alex Woodie
Rocket Software recently put the finishing touches on a range of updates for Rocket DevOps, its suite of automation, change management, and testing solutions for IBM i. The top enhancements customers will find are two-way Git integration, AI-powered anomaly detection, unit testing enhancements, and support for IBM i DevOps workflows in Microsoft Azure Pipelines.
It’s been a 2.5 years now that Rocket Software completed the rebrand from the old Aldon change management products, which it obtained with its 2011 acquisition, to Rocket DevOps. That Rocket DevOps rebrand involved taking the old Aldon Lifecycle Manager products – which helped automate the tracking and deployment of IBM i code from development into production – and adding additional components, such as Deployment Manager, Community Manager, DevOps Harmonizer, and DevOps Testing, among other new capabilities.
The software previously supported modern DevOps technologies and techniques, such as Git for a unified source code repository and Jenkins for automated deployment of code via CI/CD pipelines. With last year’s update, Rocket moved toward supporting free-format RPG, along with support for VS Code, which is rapidly becoming the most popular IDE for IBM i coders.
With last week’s update, Rocket is adding several new capabilities. Puneet Kohli, the president of the Rocket’s Application Modernization Business Unit, gave IT Jungle a briefing on the new stuff.
The addition of two-way Git integration should reduce manual effort and improve the quality of code deployed into production, Kohli said. It will do this by automatically ensuring that all of the code used in an application and stored in Git is compiled at the same time.
“Let’s say you’re developing the RPG code, but then you’re also developing the Java or React application that is sitting in the front end,” Kohli said. “Now we synchronize the changes. We are listening to the changes that are happening on the Git and then we can quickly synchronize it, so when you’re compiling the code, you’re not finding things later once it all comes together.”
Previously, that coordination of RPG and Java code, for instance, was all handled manually, Kohli. But thanks to two-way Git integration, Rocket DevOps now handles it automatically.
Rocket DevOps already supports Jenkins for CI/CD processes. But Microsoft developed its own CI/CD tooling for its cloud-based environment that is not based on open source Jekins. Rocket added support for Azure Pipelines, as the CI/CD software is known, at the request of IBM i shops that are running IBM i applications in Microsoft Azure, Kohli said.
“We’ve had requirements for customers who are deploying our applications in Azure Cloud,” Kohli said. “They want to have similar capabilities that the open systems have with the Azure Pipeline, that’s similar to Jenkins.”
Rocket is also adding AI capabilities to its various DevOps products for IBM i and System Z. The company is leveraging the large language model (LLM) capabilities of various providers, like OpenAI, to help improve the development experience of its customers.
These LLMs provide at least two new capabilities for IBM i. One of them is automated anomaly detection of coding errors. According to Kohli, using an LLM to analyze code can help customers find problems before they pop up during the compilation or deployment stage. Kohli said it’s seeing a 25 percent improvement in code quality using these technique.
“You can find compiler or configuration issues while [deploying], but we have enabled capabilities so you can find these much earlier in the cycle while you’re doing your CI/CD process with the pipeline integrations,” he said.
“The other thing we’ve added is the ability to do unit testing,” Kohli continued. “That is a capability that we’ve had in the product, but we are enhancing it so we can look at the code and make recommendation on and generate some of the unit test cases for you.”
It’s not exactly clear to what extent Rocket is using the LLMs to check IBM i code or to generate unit tests (which is a capability that already existed in Rocket DevOps), and what customers with RPG code can expect from these AI-powered capabilities. The problem stems in part from the lack of exposure that popular LLMs have had to RPG. While LLMs have been trained on mounds of COBOL that exists in the public domain, they have generally not been exposed to RPG, which generally does not exist in the open domain.
Rocket is moving forward more aggressively with integrating AI into its COBOL offerings, including its version of COBOL, called Visual COBOL. According to Kohli, Rocket is leveraging AI to deliver a code explain capability for Visual COBOL.
Eventually, Rocket wants to offer AI-powered solutions to RPG developers, Kohli said. It’s waiting to see how IBM’s Project Bob pans out, and where it can provide more value.
“That’s the direction we want to head into,” Kohli said about AI capabilities for RPG. “We definitely want to see and listen to what IBM’s doing there. They’re doing a lot of good work, so we’ll work with the community as the community enhances and provides capabilities and we’ll continue to add that to Rocket DevOps.”
RELATED STORIES
Rocket DevOps Gets Update for Free Format RPG, VS Code, and DORA
Rocket DevOps Now Supports VS Code
