stackconf 2025: A Glimpse into a Sustainable Infrastructure

At this year’s stackconf 2025 in Munich, I was particularly interested in a talk by Alessandro Vozza titled „The Sustainable Infrastructure of the Future„. In a world increasingly shaped by digital infrastructure, Alessandro posed a crucial question: How sustainable is the IT powering our modern world?

The Problem: IT’s Carbon Footprint

As a provider of virtual infrastructure and Managed Kubernetes Clusters, we’re well aware that operating hardware in data centers goes hand in hand with a significant responsibility: energy consumption and sustainability are no longer optional considerations—they’re essential, and increasingly expected by our customers.

However, for organizations running workloads in public clouds, there’s often little real influence over how that energy is sourced or the data centers are operated. Alessandro opened his talk with a brief overview of climate change, highlighting the significant role the IT industry and data centers play in contributing to global emissions.

While many cloud providers proudly advertise carbon neutrality or ambitious net zero goals, it’s just as important to take a closer look at our own workloads and ask: What can we do to reduce their environmental impact?

Practical Solutions for a Greener Infrastructure

Alessandro Vozza didn’t stop at identifying problems—he offered practical solutions for running Kubernetes workloads more sustainably:

• move workloads to greener regions or providers that use renewable energy
• use carbon-aware scheduling: schedule resource-intensive jobs during periods when renewable energy is more available
• eliminate overprovisioning by leveraging various autoscaling tools to provide scalable, efficient, and sustainable Kubernetes platforms:
  • use tools like Horizontal Pod Autoscaler (HPA), Vertical Pod Autoscaler (VPA), and the Cluster Autoscaler. A good introduction can be found in Kubernetes 101 by Daniel
  • use KEDA for event-driven workloads or when scaling based on metrics beyond CPU and memory — there’s even a carbon-aware version of KEDA that adds environmental intelligence
  • use Karpenter to get cost-efficient and dynamic node autoscaling
  • be mindful of your cluster’s energy consumption — monitor and optimize it with tools like Kepler.

Conclusion

Alessandro Vozza’s talk was a compelling reminder that while individual developers and operators might feel powerless in the face of climate change, there are levers we can pull. By making our software and infrastructure smarter, leaner, and more environmentally aware, we can all contribute to a more sustainable digital future. Fittingly, in the afternoon, Ahmad Tariq gave a short demo showing how Karpenter and KEDA can work together to run workloads more efficiently.

The other talks at today’s stackconf 2025 in Munich were equally engaging and offered a wide range of valuable insights into current trends and challenges in the infrastructure and cloud-native space. From technical deep dives to broader discussions on responsible IT practices, there was something to take away from every session.