Thanks for waiting

First of all: a big thank you to all of you who have been waiting for a new blog post. It has been a while since my last blog post — thank you for your patience and for still checking in. I appreciate everyone who reached out or simply kept reading. This week brings a proper update: a busy stretch at the Commerzbank Tech Conference in Łódź, Poland, and the kickoff of an exciting internal AI program.

Commerzbank Tech Conference in Łódź

Recently I participated in the Commerzbank Tech Conference (DTC) in Łódź, Poland. It was a great event with a strong technical audience and plenty of hands-on energy. Beyond the two presentations below, I also delivered three workshops:

  • Getting started with OpenShift 4 — fundamentals for teams moving onto the platform
  • GitOps — declarative, Git-driven operations with Argo CD and related tooling
  • Kafka — event streaming deep dive for modern architectures

The conference atmosphere was collaborative, and the workshop sessions in particular sparked good discussions about real-world adoption challenges.

Rolling Beyond DeploymentConfigs — Argo Rollouts for Next-Gen OpenShift Deployments (with Service Mesh 3)

OpenShift teams have relied on DeploymentConfigs for years. They offered automatic rollbacks, explicit triggers, lifecycle hooks, and custom deployment strategies — capabilities that plain Kubernetes Deployments did not always match. But the DeploymentConfig API is deprecated, and for new installations the native Kubernetes Deployment is the recommended path.

So what do you do when you still need progressive delivery — canary releases, fine-grained traffic shifting, metric-based promotion, automated rollbacks — without going back to DeploymentConfigs?

That is where Argo Rollouts comes in. In this session I walked through the migration story from DeploymentConfig to modern progressive delivery:

  • Why the classic DeploymentConfig + ImageStream trigger pattern is convenient but lacks auditability and is not portable
  • How a GitOps workflow with image digests (@sha256:…) gives you a fully auditable source of truth in Git
  • The mechanical steps to convert a DeploymentConfig to a Deployment (and why that alone is not the end of the story)
  • How a Rollout CRD extends the Deployment model with canary, blue-green, and analysis-driven promotion strategies
  • Live demos using the rollo sample project, progressing from rolling-update-like behavior to full canary rollouts with weighted traffic steps
  • Integration with Red Hat OpenShift Service Mesh 3 in ambient mode for production-grade traffic management without sidecar injection overhead

The key message: you do not have to give up the advanced deployment capabilities you valued in DeploymentConfigs. Argo Rollouts, combined with Service Mesh 3, gives you progressive delivery that is portable, observable, and fit for production — at no additional licensing cost beyond your OpenShift subscription.

If you are currently running DeploymentConfigs and wondering what comes next, this deck is a practical starting point. The demos are all based on open repositories you can try yourself.

The 100x Migration: Scaling Modernization with AI — Red Hat Developer Lightspeed for MTA

Modernizing legacy applications is hard. Technical debt, outdated frameworks, security risks, and the sheer cost of maintaining old stacks make every migration feel like a multi-year project. Migration Toolkit for Applications (MTA) — part of the Konveyor community and included with OpenShift subscriptions — has long helped teams assess, plan, and transform applications. Static code analysis surfaces issues with precise locations in source code, but developers still had to do the heavy lifting of writing fixes.

Red Hat Developer Lightspeed for MTA changes the economics of that work by combining static analysis with generative AI:

  • Contextual fixes, not generic chat — Developer Lightspeed uses structured migration data from MTA analysis incidents, not free-form prompts
  • Model-agnostic — connect to commercial LLMs or bring your own self-hosted model on OpenShift AI
  • RAG (Retrieval Augmented Generation) — learn from how your organization has solved similar issues in the past
  • Agentic workflows in MTA 8 — agents go beyond a single fix, identifying and addressing related changes across the codebase
  • Solution Server (MCP) — records how incidents are resolved, mines insights, and provides contextual hints for better generations
  • IDE integration — works as an extension in VS Code or OpenShift Dev Spaces, enabled via the MTA 8.0 operator

In this session, my colleague Ruben Verma and I presented the full modernization story: from the pain of legacy Java EE applications (EJB, RMI-IIOP, logging framework vulnerabilities) through MTA assessment and analysis, to live demos of Developer Lightspeed transforming code — converting JMS to reactive messaging, EJB to REST, and more. We also covered practical getting-started paths for common migrations: JBoss EAP 7 → 8, Spring Framework 5 → 6, Spring Boot 2 → 3, Java 8 → 11/17/21, and Java EE/Jakarta EE → Quarkus.

The bottom line: Developer Lightspeed is not another code generator or chatbot. It is a migration co-pilot that understands your specific technical debt because it is grounded in MTA’s analysis. That is where the 100x claim comes from — not magic, but dramatically better economics when AI is applied to structured, domain-specific modernization work.

Red Hat AI Wings Program

Something new started for me this Monday: the Red Hat AI Wings Program.

This is an internal program designed to build deep AI expertise and grow participants toward Subject Matter Expert level. The format is intense and hands-on:

  • You are teamed with a mentor — in my case, an EMEA AI Specialist Solution Architect
  • The program has three milestones, each involving AI presentations, feedback loops, and deep technical discussions
  • The goal is not surface-level AI awareness, but the kind of knowledge you need to advise customers and colleagues with confidence

The program kicked off this Monday. I am still in the early days, but it already feels like the right next step given how much of my recent work intersects with AI — from Developer Lightspeed for MTA to agentic workflows and MCP integrations. More news to come as I progress through the milestones.

What’s next

Between the Commerzbank conference, the workshop sessions, and the AI Wings kickoff, this has been a full calendar week. I will keep sharing updates here — and hopefully with shorter gaps between posts. Thanks again for reading.