i-Rays Brings Observability To IBM i Performance Problems

May 5, 2025 Alex Woodie

One of the new IBM i products that POWERUp 2025 attendees will get to check out later this month is i-Rays. Developed by the Polish software company Omnilogy, i-Rays is an observability tool that helps customers to detect performance issues with their IBM i server, investigate the causes, and then get a machine learning-based recommendation on how to fix it.

Performance issues are rare on the IBM i, which is generally a tireless transactional workhorse, but they can happen. When legitimate performance problems do crop up, it can be difficult to track down the source of the problem unless you have the right skills.

“It’s a very unique set of skills you need to have to really do this job,” said Marek Walczak, the chief commercial officer for Omnilogy. “We don’t see this kind of skills at customer sites.”

Instead of calling in an expensive consultant to diagnose and fix IBM i performance issues, the company thought it would be great if customers could have a software package that could analyze the performance problems and recommend a solution automatically. That was the genesis of the creation of i-Rays.

i-Rays does three main jobs, including monitoring of IBM i metrics, analyzing performance issues, and recommendation for fixing problems.

For monitoring, the software accesses IBM i performance information in Collection Services on a periodic basis that can be set by the customer. It collects this data – including CPU utilization, disk utilization, job waits, etc. – and moves it via JDBC to an external Linux computer that’s running the i-Rays software, where the IBM i log data is converted into the OpenTelemetry format.

Once in industry standard OTel, the IBM i performance data is analyzed using i-Rays built-in GUI, which is based on OpenSearch, an open source derivation of Elasticsearch that is distributed under an Apache 2.0 license (customers are free to use the application performance monitoring of their choice, whether its Grafana, Datadog, Dynatrace, or others). i-Rays uses proprietary machine learning algorithms to detect performance anomalies in the IBM i data at a very fine level.

“We are observing the system, how it behaves over time,” Walczak told IT Jungle. “And you can tell if the current behavior is as it should be, and as it was in the past in a similar period of time. Because the system has certain characteristics, we know that if parameters are staying within a certain range, then it’s fine.”

The third step is recommending changes that can be made when legitimate performance issues are detected. Based on the observed behavior of the box, i-Rays can automatically make certain suggestions on how to reconfigure the system or the software to get the system running back within certain parameters.

The software can make recommendations to changes on things like job runtime parameters, memory pools configuration, I/O configuration, Db2 configuration, and others. It doesn’t implement these recommendations automatically, although that could be added in the future, Walczak said. Instead, the recommendations are made as a series of system command recommendations that the administrator can take and type directly into their greenscreen console to enable, he said.

i-Rays provides observability to IBM i performance issues.

In many cases, perceived IBM i performance issues turn out not to be real performance issues. That was another reason that Walczak wanted to create i-Rays: to prevent IBM i shops from going into “panic mode” when it looks like the server is misbehaving.

“That happened at one a bank recently that really put everybody on their toes,” he said. “They started analyzing the system for a couple of days, and in a week’s time, they decided nothing happened. Nothing was wrong. But some of the metrics that they got were crazy.”

The IBM i server is such an efficient and reliable workhorse of a machine that it can’t be judged by the same metrics that govern industry-standard servers. For example, you may occasionally see IBM i’s processor utilization pegged at 100 percent, but that doesn’t necessarily mean that anything is wrong.

That’s what happened at the Polish bank. When they saw 100 percent CPU utilization, they rushed to call in consultants from IBM or Kyndryl to help fix the system. What they failed to realize is that the IBM i is built to go to 100 percent CPU utilization because its workloads are generally I/O bound as opposed to being CPU bound, Walczak said. Eventually, the bank discovered the cause of the excessive CPU utilization.

“The system was behaving normally, but there was some additional load that was created that they haven’t noticed,” he explained. “Somebody on the business side created a new channel that pushed some transactions to the system. And at the time it was populating them, it took a little longer. But the behavior of the system was perfect. Nothing wrong happened.”

Walczak, by the way, has seen some large IBM i servers in his days. Before coming back into the IBM i fold with his role at Omnilogy, he was an administrator for some of the biggest IBM i installations in the world at Polish banks – actually some of biggest in the world. He recalled spending a lot of time in the Rochester benchmarking lab, where the IBMers expressed surprise at how much stress Walczak’s company could put on the Power boxes and the OS/400 banking package from Fiserv.

i-Rays splits the gap between observability tool and performance analysis tool. There have been a handful of other products that offer some of the capability of both. But i-Rays is unique in that it can replicate the capabilities of a highly paid performance expert, but provide that expertise on a constant basis, without the billable hours.

“Our aim with the tool is kind of like giving them the top consultants in a box and doing the auditing and fixing on a daily basis,” Walczak said. “If there is a big problem, they call somebody from IBM, but we just give the consultant as a digital worker that is sitting next to the box and observing it, detecting misconfigurations, and fixing it in such a way that the system never falls into the situation where it’s really misbehaving.”

i-Rays is new to the market, but has already seen a handful of installations in Poland. Walczak will be in Anaheim, California, later this month to showcase i-Rays at COMMON’s POWERUp 2025 conference at the Disneyland Hotel and to look for a partner to help it enter the North American market.