This interview is brought to you by StormForge — automate Kubernetes rightsizing with machine learning. Smarter limits, less waste, better performance. https://ku.bz/X7ls6SKmr

Peter Zaitsev, Percona founder and Coroot Co-founder, discusses the common pitfall of over-collecting logs and metrics in Kubernetes environments. He offers two key pieces of advice:

1. Be selective in data collection
2. Utilize efficient storage technologies

Peter highlights that some organizations have "almost as much infrastructure supporting collecting and processing telemetry" as their main application, which he finds "absolutely bizarre."

He emphasizes that modern systems with data compression and vectorized processing can be "at least 10x more efficient" than older systems for managing Kubernetes logs and metrics.

Watch the full interview: https://ku.bz/GMQr-2gvT

This week on Learn Kubernetes Weekly 150:

📊 From utilization to PSI: Rethinking resource starvation monitoring in Kubernetes
🔀 Inside Intra-Node Pod Traffic in Kubernetes: How Kindnet with PTP Moves Packets
💬 The story behind the great sidecar debate
🤖 Scalable ML with Azure, Kubernetes and KEDA: Generating Inputs with 500 Pods

Read it now: https://learnkube.com/issues/150

⭐️ This newsletter is brought to you by AWS — reduce the costs of your AI infrastructure with Amazon EKS https://ku.bz/gdkVpKB3H

Karthik Ranganathan, Founder & Co-CEO at YugabyteDB, explains how to tackle tool sprawl in Kubernetes.

Instead of viewing individual tools, he recommends organizing them into functional areas. The foundation consists of:

- Infrastructure as code
- Observability and monitoring
- Resilience and uptime
- Open standards and portability

On top of this foundation, he describes how to layer resource utilization, horizontal scalability, and AI/ML automation to build robust, mission-critical platforms on Kubernetes.

Watch the full interview: https://ku.bz/4NN2xYwFh

This interview is a reaction to Sven Hans Knecht's episode https://ku.bz/SyPM8Ch43

Miguel Luna, Principal Product Manager at Elastic, explains how Observability isn't monitoring.

While monitoring addresses the known unknowns (problems you can anticipate), Observability helps surface the unknown unknowns in your infrastructure and answer questions you didn't even know you needed to ask.

Watch the full episode: https://ku.bz/WwS04jYvv

This episode is sponsored by LearnKube - get started on your Kubernetes journey through comprehensive online, in-person or remote training https://learnkube.com/training

David Sudia, Senior Product Engineer at Teleport, shares two debugging practices for Kubernetes environments.

He explains how keeping timestamped logs in a physical notebook during incidents helps track and identify solutions. David also emphasizes having a map of system layers (NAT gateways, API gateways, service meshes, etc.) and isolating each component to effectively troubleshoot complex issues.

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

This interview is a reaction to Alex Movergan's episode https://ku.bz/P5Y-NrSW5

Ángel Barrera Sánchez, Senior Platform Engineer, shares insights on the practical challenges and solutions during their transition from Jenkins to a GitOps approach using Flux.

Key takeaways include the importance of validating configurations with Jenkins, the smooth synchronization using Helm charts, and the risk management strategies to avoid production disruptions.

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

Nicholas Morey, Account Solutions Architect at Red Hat, discusses the challenges with Kubernetes secrets.

He explains how secrets stored as plain text in clusters pose security risks, leading some organizations to bypass the native secrets management entirely. Instead, these teams opt for embedding credentials directly in applications — a compromise that, while not ideal, represents a practical approach to managing sensitive information in Kubernetes environments.

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

Michael Wells, Engineering Technologist at Dell Technologies, shares his perspective on emerging tools in the Kubernetes ecosystem.

He highlights three key areas: KubeVirt for virtualization workloads; tools focused on observability, and solutions that enhance high availability and disaster recovery capabilities.

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

Ben walks through Faire's complete CI transformation, from a single Jenkins instance struggling with thousands of lines of Groovy to a distributed Buildkite system running across multiple Kubernetes clusters.

You will learn:

- How to architect CI systems that match team ownership and eliminate shared failure points across services
- Kubernetes scaling patterns for CI workloads, including multi-cluster strategies, predictive node provisioning, and handling API throttling
- Performance optimization techniques like Git mirroring, node-level caching, and spot instance management for variable CI demands

Watch (or listen to) it here: https://ku.bz/klBmzMY5-

🌟 This episode is brought to you by Testkube—the ultimate Continuous Testing Platform for Cloud Native applications. Scale fast, test continuously, and ship confidently https://ku.bz/lnxYK3s0L

With @Birthmarkb "Creatine lover" Farrell

Sai Sandeep Ogety, Director of Cloud & DevOps Engineering at Fidelity Investments, explains strategies for optimizing Kubernetes pod-level costs.

He details how organizations can reduce expenses through resource allocation, including setting appropriate CPU and memory limits, implementing auto-scaling, and using spot instances. Sandeep shares how his team achieved a 30% cost reduction through continuous monitoring with tools like KubeCost, Datadog, and Apptio, emphasizing that cost optimization should be an ongoing monthly effort rather than a one-time exercise.

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

This interview is a reaction to Kensei Nakada's episode https://ku.bz/bRd0243xQ

This interview 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

Billy Thompson, Head of Global DevOps & Platform Engineering at Akamai Technologies, shares insights on three emerging tools in the Kubernetes ecosystem.

He discusses how Crossplane has gained significant adoption in large enterprises and universities. He explores the Backstage developer portal phenomenon, noting how its community persists despite scaling challenges — similar to Kubernetes's early adoption journey. Billy also highlights the revival of container-optimized Linux distributions like Flatcar and Talos and expresses interest in testing Longhorn's ReadWriteMany storage capabilities at scale.

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

This week on Learn Kubernetes Weekly 151:

📊 Kubernetes observability from day one – mixins on Grafana, mimir and alloy
🕵️ Troubleshooting packet drops in a Kubernetes-based observability platform
🌍 How We Migrated 30+ Kubernetes Clusters to Terraform
🚪 Gateway API v1.3.0: Advancements in Request Mirroring, CORS, Gateway Merging, and Retry Budgets
🧩 Introducing Gateway API Inference Extension

Read it now: https://kube.today/issues/151

⭐️ This newsletter is brought to you by @KubeToday — a daily feed of Kubernetes news, events, jobs, announcements, and more! https://kube.today

Brian Fox, Co-Founder & CTO at Sonatype, explains why minimal containers are just the beginning of supply chain security.

He highlights that modern applications comprise 90% open source components, making dependency tracking crucial. Fox emphasizes the importance of maintaining an organization-wide Software Bill of Materials (SBOM) to enable quick responses to vulnerabilities like Log4Shell, rather than spending months tracking dependencies through email chains.

Watch the full interview: https://ku.bz/hS9sRr3sp

This interview is a reaction to Harsha Koushik's episode https://ku.bz/n_sJ04xMY

Nathan Taber, Kubernetes and Container Registries Head of Product at Amazon Web Services (AWS), explains the delicate balance between availability and efficiency in Kubernetes clusters.

He breaks down over-provisioning into two key areas: infrastructure-level optimization using Karpenter for dynamic provisioning, and application-level tuning through proper pod requests and limits. The discussion reveals how over-provisioning needs vary dramatically based on workload characteristics — from stateless applications requiring minimal overhead to stateful applications needing up to 90% extra capacity.

Watch the full interview: https://ku.bz/CYdKR23Z1

This interview is a reaction to Alexandre Souza's episode https://ku.bz/z2Vj9PBYh

Eli Birger, CTO and Co-Founder at PerfectScale, explains how observability is the foundation for controlling over-provisioning in Kubernetes clusters.

He emphasizes that while collecting metrics is straightforward, the real challenge lies in correlating data from ephemeral containers and machines. The goal isn't just gathering information but transforming raw data into actionable recommendations for immediate implementation.

Watch the full interview: https://ku.bz/6GX7MNBtS

This interview is a reaction to Alexandre Souza's episode https://ku.bz/z2Vj9PBYh

Alexandre Souza, Senior Platform Engineer at Getir, discusses the intricacies of configuring Horizontal Pod Autoscaler (HPA) in Kubernetes. He explains:

1. Three types of metrics for scaling:
- Utilization (average percentage across pods)
- Average value (of requested resources)
- Static value (specific threshold)
2. Resource-based scaling using CPU, memory, and GPU metrics
3. Scaling behaviour policies to control how pods scale up and down
4. Balancing initial requests and scaling thresholds:
- Start with conservative resource requests
- Configure HPA to scale at higher thresholds (e.g., >100% utilization)

Alex emphasizes the importance of application profiling before deployment to set appropriate resource requests and HPA configurations, ultimately helping to avoid over-provisioning in Kubernetes clusters.

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

Billy Thompson, Head of Global DevOps & Platform Engineering at Akamai Technologies, discusses how to approach building internal platforms.

He emphasizes that organizations should first answer "why do you need an internal platform?" and evaluate whether it will reduce existing toil. The discussion covers essential considerations including:

- Assessing if your organization has reached the right scale for platform engineering
- Securing bottom-up adoption from SREs and development teams instead of top-down mandates
- Evaluating how vendor dependencies impact long-term flexibility
- Finding the balance between building from scratch and using managed solutions

Billy advocates for choosing solutions that are flexible and portable while avoiding additional operational complexity.

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

This interview is a reaction to Ángel Barrera Sánchez's episode https://ku.bz/-5QbzQXJg

Eric Jalal, Kubernetes Architect, explains how Kubernetes builds upon Linux APIs while adding its own design principles. He highlights:

- Wrapping APIs is complex, requiring extensive testing and experience
- Kubernetes incorporates best practices from Google's data centers
- It's designed for high uptime, flexibility, and modularity
- The architecture considers business needs, organizational structure, and CI/CD processes
- Kubernetes combines software development and operations perspectives

Eric emphasizes that Kubernetes is more than just a Linux wrapper - it's a purposefully designed system for modern, complex environments.

Watch the full episode: https://ku.bz/-jCTfgqRC