In this episode, recorded live at KubeCon, Ronald and Jan talk with Jannis Relakis and Michael Seiwald-McCarty, both senior platform engineers at Celonis. Celonis manages over 150 Kubernetes clusters across GKE, AKS, and EKS, but it wasn't always that clean. They started with six different Kubernetes flavors, including Gardener, K-Ops, and OpenShift (both Rosa and ARO), spread across multiple cloud providers.

In their KubeCon talk "No Shame in Just Paying," they shared how they tackled this consolidation project: migrating all workloads to three standardized, fully managed Kubernetes distributions. Key topics include their self-built cross-cluster connectivity tool called "Wormhole" (powered by Envoy's dynamic forward proxy), RabbitMQ federation for seamless message queue migration, and how they used Karpenter and Cilium to align node and network management across clouds.

They also get candid about what went wrong: an accidental ArgoCD sync that caused a 10-minute full environment outage, the pain of "snowflake" environments (including one requiring full HIPAA compliance with Istio mTLS), and the constant fight against scope creep that threatened to derail the entire project.

The episode closes with a forward-looking discussion on FinOps, resource rightsizing, the future of VPA, and whether Kubernetes and serverless can ever truly converge.Powered by ACC ICT

Stuur ons een bericht.

ACC ICT Specialist in IT-CONTINUÏTEIT
Bedrijfskritische applicaties én data veilig beschikbaar, onafhankelijk van derden, altijd en overal

Support the show

Like and subscribe! It helps out a lot.

You can also find us on:
De Nederlandse Kubernetes Podcast - YouTube
Nederlandse Kubernetes Podcast (@k8spodcast.nl) | TikTok
De Nederlandse Kubernetes Podcast

Where can you meet us:
Events

This Podcast is powered by:
ACC ICT - IT-Continuïteit voor Bedrijfskritische Applicaties | ACC ICT

In this episode, Ronald and Jan sit down with Luigi Nardi, founder and CEO of DB tune, at KubeCon. Luigi brings a rare mix of academic depth (PhD in computer science, postdocs at Imperial College London and Stanford, professor at Lund University) and startup pragmatism. The conversation digs into why database tuning is fundamentally a combinatorial optimization problem that humans aren't wired to solve well, and why AI is uniquely suited for it.

DB tune focuses entirely on Postgres and deploys a narrow, production-safe AI agent that reads performance metrics and iteratively adjusts server parameters (GUCs) until the system converges on an optimal configuration. No LLMs, no hallucinations — just purpose-built ML that operates in a closed feedback loop. The agent integrates with AWS RDS, Aurora, Azure Flexible Server, Google Cloud SQL, and Cloud Native PG (the Kubernetes Postgres operator).

Luigi shares a standout story: a water management company ran the agent on their production system — with a hospital's water supply on the line — and achieved a 2.5x performance improvement in just a few hours. He also explains how tuning isn't a one-time exercise: cloud workloads change, hardware scales up and down, and DB tune's model called "Newton" was specifically engineered to prevent unstable, oscillating parameter changes. The episode closes with a compelling FinOps angle: tuning doesn't just make your database faster, it can also shrink your instance size and cut infrastructure costs — a perfect fit for the Kubernetes-native world.

Powered by ACC ICT

Stuur ons een bericht.

ACC ICT Specialist in IT-CONTINUÏTEIT
Bedrijfskritische applicaties én data veilig beschikbaar, onafhankelijk van derden, altijd en overal

Support the show

Like and subscribe! It helps out a lot.

You can also find us on:
De Nederlandse Kubernetes Podcast - YouTube
Nederlandse Kubernetes Podcast (@k8spodcast.nl) | TikTok
De Nederlandse Kubernetes Podcast

Where can you meet us:
Events

This Podcast is powered by:
ACC ICT - IT-Continuïteit voor Bedrijfskritische Applicaties | ACC ICT

In this episode, hosts Ronald and Jan are joined at KubeCon by two guests from Red Hat: Brian Stevens, AI CTO and one of the original architects behind the creation of Kubernetes and the CNCF, and Rob Shaw, co-lead of the vLLM project and maintainer of LMD.

Brian shares the remarkable backstory of how Kubernetes came to be open source, including how Red Hat negotiated a single committer seat before agreeing to be a launch partner, and how he later pushed Google to contribute Kubernetes to the newly formed CNCF rather than keeping it proprietary like TensorFlow.

Rob explains what an inference runtime actually is: the critical piece of software that takes an abstract AI model and runs it as efficiently as possible on a GPU or other accelerator — handling everything from CUDA-level kernel optimization to memory management and concurrent request scheduling. vLLM serves as a "Rosetta Stone" between the ever-growing zoo of models (Llama, DeepSeek, Mistral, Qwen, Nvidia Nemotron) and accelerators (Nvidia, AMD, Intel, Google TPUs).

The conversation covers model compression and quantization how techniques like 4-bit precision can deliver 2x hardware efficiency gains while preserving 99%+ model accuracy. Brian and Rob also address the "big model vs. many small models" debate, recommending to always start with the largest capable model to validate a use case before optimizing down.

Looking ahead, both guests see inference as potentially the single largest workload ever run on Kubernetes, and position LMD (now contributed to the CNCF) as the distributed inference layer that will make this possible across heterogeneous accelerator environments preventing enterprises from ending up with 42 incompatible AI stacks.
The episode closes with a discussion on AI slop, human-in-the-loop thinking, and the future of Kubernetes as the universal platform for running AI agents at scale.

Powered by @acc-ict ​

Stuur ons een bericht.

ACC ICT Specialist in IT-CONTINUÏTEIT
Bedrijfskritische applicaties én data veilig beschikbaar, onafhankelijk van derden, altijd en overal

Support the show

Like and subscribe! It helps out a lot.

You can also find us on:
De Nederlandse Kubernetes Podcast - YouTube
Nederlandse Kubernetes Podcast (@k8spodcast.nl) | TikTok
De Nederlandse Kubernetes Podcast

Where can you meet us:
Events

This Podcast is powered by:
ACC ICT - IT-Continuïteit voor Bedrijfskritische Applicaties | ACC ICT