Gabriele Bartolini, VP of Cloud Native at EDB, explains why the "start small" approach for Postgres on Kubernetes is unthinkable in modern deployments.

He recommends using operators like CloudNativePG from day one to bypass common challenges. The key insight is that there are no performance penalties when running Postgres in Kubernetes compared to VMs or bare metal — the focus should be on treating Postgres as a first-class citizen in the cluster rather than experimenting with minimal setups.

Watch the full interview: https://ku.bz/xsBRT9m0L

This interview is a reaction to David Pech's episode https://ku.bz/rGMF2ktdb

Ángel Barrera Sánchez, Senior Platform Engineer, shares insights on the pivotal transition to a GitOps approach at Adidas.

Angel emphasizes the foundational mindset of treating everything as code, a principle established by the initial team in 2018 that guided the project's evolution.

Watch the full episode: https://ku.bz/-5QbzQXJg

Billy Thompson, Head of Global DevOps & Platform Engineering at Akamai Technologies, shares his perspective on learning Kubernetes effectively.

He advocates going beyond "learning by doing" to "learning by fixing" — deliberately breaking systems and solving problems. This method engages multiple parts of the brain, from problem-solving to creativity, making lessons stick better than following tutorials. Billy explains why troubleshooting is a blessing and suggests creating controlled environments for breaking and fixing things, rather than experimenting in production.

Watch the full interview: https://ku.bz/d-s2179ff

This interview is a reaction to Mathias Pius's episode https://ku.bz/WxLPC_Wlb

Alexandre Souza, Senior Platform Engineer at Getir, discusses strategies to manage over-provisioning in Kubernetes clusters. He shares:

- The motivation behind his article on cost reduction initiatives
- How to optimize CPU and memory usage in Kubernetes environments
- The importance of setting conservative resource requests based on actual application needs
- Using monitoring tools to identify real resource consumption patterns

Alex emphasizes that resource requests are crucial as they determine pod scheduling and serve as the chargeable unit for cloud providers. He stresses the need to be "very conservative" when choosing metrics and values for resource allocation.

Watch the full episode: https://ku.bz/z2Vj9PBYh

Sai Sandeep Ogety, Director of Cloud & DevOps Engineering at Fidelity Investments, discusses the tools shaping the future of Kubernetes operations.

He identifies Lens and OpenTelemetry as key solutions for platform streaming and observability. In his experience, the most pressing challenge is cluster management across cloud providers—working with clusters spread across Google, Amazon, and Azure requires better tooling for unified access and control. Sandeep emphasizes the need for tools that enable single-command cluster access to streamline daily operations in multi-cloud environments.

Watch the full interview: https://ku.bz/-kMN8QDYN

This episode is brought to you by Akamai Cloud — build and deliver low-latency, edge native applications on the world's most distributed cloud computing platform. https://ku.bz/vpgt3zXMs

Nicholas Morey, Account Solutions Architect at Red Hat, discusses the different approaches to detecting vulnerable secrets in Kubernetes clusters.

While some tools use regex-based detection and others rely on entropy-based analysis, Nicholas advocates using External Secrets to integrate with dedicated secret management systems rather than relying on Kubernetes' native secrets.

Watch the full interview: https://ku.bz/Rw2DmxNn-

Emin Laletović shares his experience debugging a critical production issue in a Kubernetes environment running a Go application. He details:

1. Investigating a specific API endpoint failing with a 502 bad gateway error.
2. The systematic approach to isolate the problem.
3. Discovering that the pod restarts every time the problematic endpoint is called.

This case study highlights the complexities of production debugging in containerized environments and the importance of a methodical troubleshooting process.

Watch the full episode: https://ku.bz/7fnF-tJ8R

Michael Wells, Engineering Technologist at Dell Technologies, discusses how running databases on Kubernetes has evolved from being discouraged to becoming a recommended practice.

He explains how Kubernetes maturity and features like persistent volumes enable stateful applications to run reliably. Michael highlights two main benefits: reduced platform overhead and faster database scaling in cloud environments, while emphasizing the importance of adapting traditional database architectures for containerized deployments.

Watch the full interview: https://ku.bz/LthjkdpPg

This interview is a reaction to Steven Sklar's episode https://ku.bz/-k6hrcpV-

Billy Thompson, Head of Global DevOps & Platform Engineering at Akamai Technologies, shares his least favourite Kubernetes feature: a recurring challenge in Cert Manager's HTTP validation when using proxy mode with load balancers.

He explains how this widespread issue impacts observability and security features in Kubernetes clusters, particularly when monitoring web traffic logs.

Watch the full interview: https://ku.bz/bh07VCK23

Stephan Schwarz walks through his systematic approach to performance testing Kubernetes applications.

You will learn:

- Why shared Kubernetes components skew results and how ingress controllers, service meshes, etc. create testing challenges that require careful consideration of the entire request chain
- Practical approaches to HPA configuration, including how to account for scaling latency and planning for spare capacity based on your SLA requirements
- The role of observability tools like OpenTelemetry in production environments where load testing isn't feasible, and how distributed tracing helps isolate performance bottlenecks across interdependent services

Watch (or listen to) it here: https://ku.bz/yY-FnmGfH

🌟 This episode is brought to you by Learnk8s — get started on your Kubernetes journey through comprehensive online, in-person or remote training https://learnk8s.io/training

With @Birthmarkb "Not Open Source" Farrell

Jason (Jay) Smith, Senior Cloud Customer Engineer at Google, discusses how the infrastructure as code paradigm needs to evolve by embracing a true "code-first" approach.

He advocates for representing Kubernetes objects as native programming language objects (like Python), making the platform more accessible to developers and data scientists who already work in these languages. Jay highlights how this approach could address the fact that 50% of Kubernetes users believe the platform could be easier to use, and explains that emerging tools that transform code into YAML or Terraform HCL are steps in the right direction for platform engineering that meets developers where they are.

Watch the full interview: https://ku.bz/B1D6ZWFw1

This interview is a reaction to Brian Grant's episode https://ku.bz/_ZLj6ZV-9

This episode is brought to you by Akamai Cloud — build and deliver low-latency, edge native applications on the world's most distributed cloud computing platform. https://ku.bz/vpgt3zXMs

Eyal Bukchin, CTO & Co-Founder at MetalBear, highlights three emerging Kubernetes tools worth watching.

He discusses GlassKube for tackling on-premises software management for clients, Robusta for incident response in complex production environments, and LocalStack for enhancing local development workflows. As the creator of Mirrord, Eyal emphasizes tools that improve the developer experience and solve real operational challenges in Kubernetes ecosystems.

Watch the full interview: https://ku.bz/H5RPCdFCy

This week on Learn Kubernetes Weekly 133:

🧙‍♀️ A journey of writing my own Kubernetes
📊 Scaling Virtual Machines in Kubernetes Clusters: Insights for Kubernetes Applications
🕵️ Exploring the Kubernetes API Server Proxy
🥋 CVE-2024–10220: Attack and Defense
👧 Exploit me, baby, one more time: command injection in Kubernetes Log Query

Read it now: https://learnk8s.io/issues/133

⭐️ This newsletter is sponsored by Fairwinds — expert-led, fully managed Kubernetes that frees your engineers from infrastructure headaches and puts you on the fast track to production-grade success https://ku.bz/sSRQp8xDH

Yasmin Rajabi, CSO at CloudBolt Software, explains how their platform solves three key optimization challenges in Kubernetes environments.

She details their machine learning approach to node-level optimization that analyzes workload profiles and sets node affinities to reduce waste. Yasmin highlights their Java-specific optimization capabilities that collect heap garbage collection metrics to safely recommend memory settings for Java workloads.

Watch the interview: https://ku.bz/9z8DF_FxJ

Read the announcement: https://ku.bz/yTfnmb7GQ

Petar Cvitanović, Co-founder & CTO of Cyclops UI, discusses three emerging Kubernetes tools he's watching closely.

He highlights tools that treat Kubernetes as "a centralized place to configure all of your software" rather than just a container runtime. Petar specifically mentions Crossplane for its declarative configuration approach, KRO from AWS as a simpler Crossplane alternative, and Backstage for providing developers with autonomy over their applications and systems.

Watch the full interview: https://ku.bz/WnFG9mVvg

This episode is brought to you by Akamai Cloud — build and deliver low-latency, edge native applications on the world's most distributed cloud computing platform. https://ku.bz/vpgt3zXMs

Learn about Crossplane vs Terraform and the key differences between these infrastructure management tools with Lior Lieberman, SRE at Google.

Watch the full interview: https://ku.bz/xTk6Wswjd

This interview is a reaction to Dan Garfield's episode https://ku.bz/m3YNgCh1W

This episode is brought to you by Akamai Cloud — build and deliver low-latency, edge native applications on the world's most distributed cloud computing platform. https://ku.bz/vpgt3zXMs

Nick Nikitas Senior Platform Engineer at Blueground discusses the scalability challenges they faced with static pre-production environments in Kubernetes.

With only 1-2 dedicated environments per engineering team, they encountered three critical issues: teams informally sharing environments without coordination, miscommunication during debugging sessions, and environments becoming feature-specific when multiple developers worked simultaneously.

Watch the full episode: https://ku.bz/tt4VFslxD