i-Rays Performance Analyzer Now Ready for Prime Time, Omniology Says
May 18, 2026 Alex Woodie
Companies that want to address performance issues on their IBM i applications may want to check out the i-Rays software from Omnilogy. The product has matured considerably since it was introduced to the U.S. market last year, and it now has the backing of prominent IBM i performance expert Dawn May.
When we first wrote about i-Rays back in May 2025, the offering was brand new to the American market. Marek Walczak, the chief commercial officer for the Polish firm Omniology and the general manager of i-Rays, attended the COMMON POWERUp 2025 conference in Anaheim, California, to test the American waters for the new product, which had been installed in just a few IBM i sites in Poland.
One year later, the situation has changed. Not only is the i-Rays product more mature and running in production with IBM i customers, but Omnilogy has also strengthened the North American sales and support organization behind it.
“A year ago, i-Rays was a product, but it was more a promise,” Walczak told IT Jungle at the recent POWERUp conference in New Orleans, Louisiana. “Now it’s mature. We are fitting into the market needs.”
And the market needs are great, according to Walczak, who has been an IBM i professional for much of his career and helped start COMMON Poland in 1997. When Walczak started his IBM i career at a Polish bank in 1992, OS/400 applications were not changed very often and had little interaction with outside components. That minimized the potential for performance problems.
Fast-forward to 2026, and things are much different. IBM i applications, such as core banking systems, are connected to all sorts of additional applications, including Web and mobile front-ends, other backend Java products, etc. And it’s not uncommon for IBM i applications to be modified once or even twice a day.
Today, the risk of something going wrong due to an update to the IBM i app itself or one of the many products it’s integrated with is much greater. However, the tools for detecting those performance issues haven’t changed much. These performance issues can be detected and acted upon before it’s too late, but it requires having people with the right set of skills to understand what the IBM i performance data means.
That’s where i-Rays comes in. The product is based on OpenSearch, an open version of Elasticsearch, and collects data in the OpenTelemetery format. It includes a series of proprietary machine learning algorithms to detect patterns hidden in IBM i performance data, which customers can see in Grafana or other open source visualization tools.
It is notoriously difficult to find patterns hidden in any time-series data, let alone data coming from a system that few mainstream observability tools really understand. That is the key capability that i-Rays delivers. While observability vendors like Dynatrace support IBM i, they lack the deep understanding of the IBM i platform and its architecture.
“They approach IBM i the same way as they approach Windows and Linux, and it’s no good, because IBM i is so unique,” Walczak said.
Omniology tackled the skills gap by bringing together two individuals, one who is an expert in observability and one who is an expert in IBM i, according to Walczak. It took them six months to develop a common understanding and a common language of how to tackle observability on IBM i.
The result was an i-Rays product that has a built-in affinity for IBM i and how it runs. For example, having a CPU load of 100 percent could be a sign of a serious problem on industry-standard servers, but it could be perfectly fine on the IBM i server.
i-Rays uses time-series algorithms to detect subtle changes in IBM i performance data.
“It’s not just burned out. It’s just busy,” Walczak said. “That’s why what we do is we recognize the load patterns. And we can tell you, okay, your load patterns are different now.”
The key is to detect when subtle changes in load patterns are an indicator that something is seriously wrong. This requires a deeper understanding of what’s going on within IBM i applications, which i-Rays does through the use of markers.
Walczak shared the story of a recent customer that was running i-Rays in proof-of-concept mode. The bank implemented some changes to an RPG program after testing on Friday evening. On Saturday morning, the CPU load was sitting at 5 percent. Normally, that wouldn’t raise any red flags, but i-Rays recognized that something was wrong.
On Monday morning, the bank opened up to the horror of its payment processing system being down, putting a stop to billions of dollars’ worth of transactions. Since i-Rays was not fully implemented, the suggestion that it generated – the so-called “calming” commands that i-Rays can automatically generate to alleviate performance issues – were not executed. The source of the problem wasn’t the RPG change directly; it was a problem in the multi-threading used by the data replication software that the bank relied on for high availability.
You can never fully rely on testing after you make changes to a production system, Walczak said. “You can only pray that on Monday, when the load goes up, that you will survive,” he said. “But we can tell you, because we need only 5 percent of the load to recognize patterns. We can tell you your program is now running differently than before the change.”
Omniology’s approach has captured the attention of Dawn May, the leading IBM i performance expert in the world. May, who received the COMMON President’s Award at the recent POWERUp conference, says she likes what i-Rays is doing.
“They are using the system performance data (i.e., Collection Services) as the key data source. This has a very rich metric set, and good efficiency with little overhead for accessing the data,” May tells IT Jungle. “This is really important since I don’t like to see performance tooling where the tool becomes one of the largest consumers of CPU.”
Mainstream observability tool vendors usually require a program to be running directly on the IBM i. Walczak said that can be a dealbreaker in regulated industries, and is the reason that i-Rays does not require any software to be installed on IBM i.
May continued: “The AI/ML they have built allows the tool to identify potential issues and possible solutions. With the complexity of IBM i, it can be difficult for newer system administrators to know all the potential places to look for problems, so it’s great to see AI used for tools on the administration side.”
May, who is working as a consultant for Omniology, wrote two articles on two topics where she sees i-Rays adding value. You can read them at https://dawnmayi.com/2026/04/23/an-alternative-dynatrace-extension-for-ibm-i/ and https://dawnmayi.com/2026/04/13/time-to-forget-set-and-forget/.
In the meantime, Walczak has garnered a new business partner to represent i-Rays in the United States. For more information, see i-rays.com.
Editor’s note: This story was corrected. COMMON Poland was founded in 1997. IT Jungle regrets the error.
