Zbyněk Roubalík, CTO Founder @ Kedify, shares his perspective on effective auto-scaling strategies for Kubernetes.

He breaks down auto-scaling into two essential levels: cluster-level scaling for nodes and application-level scaling. He stresses that successful auto-scaling requires choosing the right metrics beyond just CPU and memory to make scaling decisions.

Watch the full interview: https://ku.bz/k0xRzNBVd

This interview is a reaction to Thibault Jamet's episode https://ku.bz/rf1pbWXdN

ingress2gateway is a CLI that reads Kubernetes Ingress resources and outputs equivalent Gateway API manifests.

It maps hosts, paths, and routing rules directly, allowing you to adopt Gateway API resources instead of legacy Ingress.

More: https://ku.bz/Ymkzv38cM

Calin Florescu explains how inefficient Helm templates and inconsistent service configurations translate into business costs.

He emphasizes that time is the critical factor — delays in deployment processes directly impact the product feedback cycle. While engineers might dismiss YAML formatting or template inconsistencies as minor issues, these technical inefficiencies can significantly extend the time between releases and slow down the time-to-market, resulting in tangible financial losses for the business.

The revised version better uses markdown formatting while maintaining the core message.

Watch the full episode: https://kube.fmhttps://ku.bz/mcPtH5395

Sveltos installs as a controller in a management cluster, deploying add-ons and policies (Helm charts, Kustomize, raw YAML) to target clusters by label selectors and sync rules, automating multi-cluster resource management and compliance.

More: https://ku.bz/j_ZZTyYqy

SchemaHero is a Kubernetes operator that manages database schema changes declaratively using custom resources.

It applies, versions, and updates schemas automatically for PostgreSQL, MySQL, and other databases in CI/CD pipelines.

More: https://ku.bz/BNMQt-WGF

This article explains how to prevent system failures caused by imbalanced scaling in Kubernetes using KEDA with proper trigger configurations and scaling strategies.

More: https://ku.bz/BYr3WgBhp

This week's 6 best Kubernetes leadership jobs are:

Site Reliability Engineer with WP Engine
💰 $99.18K to $148.77K a year
👨‍💻 Remote from Poland
→ https://ku.bz/3RZRJPRR1

DevSecOps Engineer with OpenAI
💰 $364.5K to $490K a year
👨‍💻 Remote from the United States of America
→ https://ku.bz/NXd17JHfV

Software Engineer with 2100 NVIDIA USA
💰 $272K to $425.5K a year
👨‍💻 Remote from the United States of America
→ https://ku.bz/MM_GYJjMZ

Platform Engineer with OKX
💰 $268.33K to $375.67K a year
🌎 Fully remote
→ https://ku.bz/7nDz6kmPm

Software Engineer with Reddit
💰 $260.8K to $365.1K a year
🌎 Fully remote
→ https://ku.bz/yfQxGWT6D

👉 Browse 1282 jobs on Kube Careers https://kube.careers

GitOps Bridge is a community project that aims to showcase best practices and patterns for bridging the process of creating a Kubernetes cluster and subsequently managing everything through GitOps.

More: https://ku.bz/_9TTyPgK3

Sai Sandeep Ogety, Director of Cloud & DevOps Engineering at Fidelity Investments, discusses the limitations of cross-cloud integration in Kubernetes.

He explains how connecting an AWS Kubernetes cluster to Azure storage requires complex workarounds due to missing native features. The discussion focuses on two main challenges: storage access across different cloud providers and network connectivity between cloud and on-premises environments. Sandeep advocates for cloud-agnostic features in Kubernetes that would simplify resource sharing and communication in hybrid environments.

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

Upjet lets you generate Crossplane providers by combining a codegen pipeline, a generic CRD reconciler, a doc scraper, and migration tooling.

More: https://ku.bz/fsXcFNy_p

Helm Chart OCI Proxy transparently proxies and transforms Helm Charts as OCI artefacts.

You can use it to address any public Chart Repository-styled Helm Chart as an OCI-styled artefact.

More: https://ku.bz/KJks89SqP

John McBride, VP of Infrastructure and AI Engineering at The Linux Foundation, explains why Kubernetes was the ideal platform for deploying VLLM to run AI inference workloads at scale.

He highlights two key technical advantages: dynamic GPU scaling capabilities and the ability to run internal Kubernetes services behind Kubernetes DNS and load balancing without third-party dependencies.

John emphasizes that despite the operational burden, the ability to immediately fix issues rather than depend on potentially volatile third-party services was worth the tradeoff. He concludes with his conviction that "Kubernetes is the platform of the future for AI, ML" particularly when operating at scale with cloud-based GPU resources.

Watch the full episode: https://kube.fmhttps://ku.bz/wP6bTlrFs

KCL allows developers to create modular, scalable, and stable configurations.

Key features include constraint-based records, functional language, schema modeling, and automation APIs, ideal for cloud-native environments and platform engineering.

More: https://ku.bz/55b96R-MD

This week's 6 best Kubernetes architect vacancies are:

Solution Architect with Chainguard
💰 $12.95M to $14.97M a year
👨‍💻 Remote from Australia
→ https://ku.bz/g8LNJ6CT5

Solution Architect with NVIDIA
💰 $308K to $471.5K a year
🏠 From the office in Santa Clara, CA / Champaign, IL, USA
→ https://ku.bz/VJNQljZmd

Solution Architect with Caylent
💰 $250K to $450K a year
👨‍💻 Remote from
→ https://ku.bz/fprtkLknc

Software Architect with Toshiba Global Commerce Solutions - External
💰 $250K to $280K a year
👨‍💻 Remote from
→ https://ku.bz/28YdKD1Fp

Software Architect with Conga
💰 $197.46K to $315.94K a year
👨‍💻 Remote from the United States of America, Canada
→ https://ku.bz/28QRgfPW4

👉 Browse 1291 jobs on Kube Careers https://kube.careers

This week on Learn Kubernetes Weekly 167:

⚖️ Kubernetes & KEDA: Avoiding System Failures from Imbalanced Scaling
🔐 Why DevOps should Sec: making a case for DevOps Engineers to transition to DevSecOps
🌐 Optimizing Pod IP Allocation in AWS EKS with Amazon VPC CNI Prefix Delegation
🎮 GPU Starvation in Kubernetes: How Dynamic MIG Partitioning Saved Our GPU Budget
🔄 Migrating from F5 NGINX ingress controller to the F5 NGINX gateway fabric

Read it now: https://kube.today/issues/167

⭐️ This newsletter is brought to you by LearnKube — master Kubernetes with hands-on training designed for engineers who want to learn the smart way https://ku.bz/hypSbyc-V

Headlamp is an easy-to-use and extensible Kubernetes web UI providing a user-friendly interface for managing and debugging Kubernetes clusters.

More: https://ku.bz/1jVt_89N-

Capsule is a Kubernetes operator that enables multi-tenancy by grouping namespaces into Tenant CRDs.

It enforces isolation via RBAC, NetworkPolicies, and ResourceQuotas, allowing self-service namespace creation within policy-defined boundaries.

More: https://ku.bz/k1trm6F6g

flux-operator simplifies the configuration of Flux multi-tenancy lockdown, sharding, horizontal and vertical scaling, persistent storage, and allows fine-tuning the Flux controllers with Kustomize patches.

More: https://ku.bz/cKqych-t6

Kusion is a platform orchestrator for IDPs that enables declarative infrastructure and application delivery across environments.

It compiles specs into IaC plans and deploys via Terraform, Helm, or Kubernetes APIs.

More: https://ku.bz/7-zQDyFZv

David Pech, Staff Cloud Ops Engineer, breaks down a legacy PHP application's architecture that exemplified everything modern cloud-native development tries to avoid. He describes a custom-built PHP system without any framework where developers violated MVC principles by embedding logic in views and relied entirely on function-based programming instead of object-oriented design.

David explains how the team's "throw it over the wall" mentality left developers with zero operational feedback, resulting in applications with virtually no observability beyond basic error logging in Postgres.

This real-world example illustrates how legacy systems accumulate technical debt through poor architectural decisions and highlights why operational experience and feedback loops are crucial for building maintainable applications.

Watch the full episode: https://ku.bz/_MWX5m6G_

pv-migrate is a CLI tool/kubectl plugin to easily migrate the contents of one Kubernetes PersistentVolumeClaim to another.

More: https://ku.bz/32y60KXjl