Learn Kubernetes Weekly issue 180 · 22 Apr 2026
Distributed LLM Inference Challenges, Model Serving with Ray, Lazy Image Pulling, eBPF Based Bandwidth Limiting, Slurm on Kubernetes
This newsletter is brought to you by Portworx. Automate, protect, and unify data for modern applications across on-premises, public, and hybrid cloud environments.
Articles
Hidden Infrastructure Challenges in Distributed LLM Inference on Kubernetes
substack.com
This article explains why distributed LLM inference on Kubernetes is so hard to get right: your GPUs and network cards need to be physically close on the same PCIe switch, but Kubernetes pairs them at random and kills your RDMA performance.
How Kubernetes Storage Actually Works
portworx.com
This article explains how storage works in Kubernetes and covers PersistentVolumes, StorageClasses, CSI drivers, snapshots, backup and disaster recovery for stateful workloads.
sponsored
Simplifying Model Serving with Kubernetes and Ray: Inside DoubleVerify’s ML Platform
medium.com
This case study shows how DoubleVerify built a Kubernetes and Ray serving platform to deploy and scale ML models in production.
It also covers RayService wrapped with Helm, fault tolerance with external Redis, and platform gains like 30% lower GPU cost.
Lazy-pulling container images: a deep dive into OCI seekability
blog.zmalik.dev
This article covers:
- why OCI container layers resist random access due to DEFLATE dependency chains,
- benchmarks eStargz, SOCI, Nydus, and cloud-managed lazy-pulling approaches,
- how FUSE-based lazy pulling shifts cost from pull to runtime.
Building eBPF-Based Bandwidth Limiting in AWS Network Policy Agent — Why Vibe Coding Isn’t Enough
medium.com
This article walks you through building EDT-based eBPF bandwidth limiting in the AWS Network Policy Agent, showing where AI-generated code silently broke and how domain knowledge caught each bug.
Slurm on Kubernetes (SUNK): Modernizing HPC and AI workload management
medium.com
This article explains how Slurm on Kubernetes combines Slurm job scheduling with Kubernetes orchestration so AI and HPC teams can modernize GPU-heavy infrastructure without forcing researchers into raw Kubernetes workflows.
The Voice of Kubernetes Report 2026
Where is Kubernetes headed in 2026?
519 infrastructure teams share what workloads they're running, where backup and DR is still the biggest gap, and what the next 5 years look like.
Tutorials
[Webinar]Virtualization Reimagined: How to Escape Your Rising VM Costs
portworx.com
This webinar by Portworx covers how Everpure migrated 5,000+ VMs onto Kubernetes using KubeVirt and Portworx to cut legacy virtualization costs and unify VM and container workloads.
sponsored
Hardware-backed TLS certificates with cert-manager and yubihsm 2
charles.dev
This tutorial teaches how to build a cert-manager external issuer that uses a YubiHSM 2 to sign TLS certificates via Go's crypto.Signer interface.
Mastering KEDA on GKE: A Deep Dive into Event-Driven Autoscaling
saeed.hashnode.dev
This tutorial explains how to use KEDA on GKE to autoscale workloads based on event-driven signals rather than just CPU or memory.
Freezing Spark Drivers to Zero Resources and Waking Them in 300 Milliseconds
medium.com
This article explains how Spark Connect, CRIU, and ZeroPod can freeze idle Spark drivers to near-zero resources and restore full session state in about 300 milliseconds on Kubernetes.
ing-switch: Migrate from Ingress NGINX to Traefik or Gateway API in Minutes, Not Days
blog.kubesimplify.com
This article introduces ing-switch, a tool that scans Kubernetes ingress resources and helps teams migrate from Ingress NGINX to Traefik or Gateway API by mapping annotations and showing compatibility gaps.
Kelsey Hightower, Eric Abercrombie, and Julius Payne II reflect on life after achievement, entering the Kubernetes world for the first time, and how music, creativity, and lived experience shape the way they think about technology.
In this interview:
- Why fundamentals, patience, and repetition still matter more than shortcuts
- How Kubernetes, community, and confidence intersect for people entering cloud-native work
- What hip-hop, production, and storytelling can teach us about ownership, authenticity, and finding your voice
Kubernetes jobs
Machine Learning Engineer with xAI
Salary: $135K to $393.25K a year
Location: based in the office in Palo Alto, CA, USA
Tech stack: Kubernetes, Kubernetes, Ray, Spark, XLA, Jax, LLMs, PyTorch, C++, Python
Software Engineer with Inetum
Salary: $23.76K to $125.4K a year
Location: based in the office in Lima, PE
Tech stack: Kubernetes, Kubernetes, AKS, Azure, On-premise, Application Insights, GitHub, Git, Azure DevOps, Azure Keyvault
Software Engineer with Devoteam
Salary: $126K to $275K a year
Location: based in the office in Nantes, FR
Tech stack: Kubernetes, Kubernetes, Google Cloud, Microsoft Azure, AWS, Docker, CI/CD, cybersecurity, MongoDB, DynamoDB
Solution Architect with Veeam Software
Salary: $84.6K to $346.5K a year
Location: remote from
Tech stack: Kubernetes, Kubernetes, AWS, Azure, Docker, GCP, On-premise, compliance, ISO 27001, NIST
Salary: US$96.3K to US$286K a year
Location: based in the office in Zürich, CH
Tech stack: Kubernetes, Kubernetes, Cloud, Docker, On-premise, OpenShift, alerting, monitoring, logging, tracing
Discover more Kubernetes jobs on Kube Careers →
Code & tools
github.com/yindia
RootCause is a local first MCP server for Kubernetes that turns natural language into evidence backed incident analysis, safe operation checks, and ecosystem diagnostics for tools like Argo CD, Flux, Cilium, and Helm.
Warden for Identity-Based Access Control for AI Agents and Kubernetes Workloads
github.com/stephnangue
Warden is an open source runtime access gateway that lets AI agents, pods, pipelines, and services use identity-based policies to reach cloud APIs, databases, and storage without storing long-lived credentials.
GreenKube: carbon and cost visibility for Kubernetes
github.com/GreenKubeCloud
GreenKube is an open source platform that measures Kubernetes workload energy use, estimates CO2e emissions, and gives optimization recommendations so teams can reduce cloud cost and carbon impact.
github.com/vllm-project
AIBrix is a Kubernetes-native GenAI inference infrastructure toolkit from the vLLM project, with LLM-aware routing, distributed KV cache, LoRA management, and an app-tailored autoscaler for vLLM workloads.
github.com/FairwindsOps
Pluto scans Kubernetes manifests, Helm charts, and live Helm releases to find deprecated or removed API versions before upgrades break workloads.
Other interesting projects:
Upcoming Kubernetes events
Apr
23
Online workshop organized by LearnKube.
This is a virtual event
This event requires an entrance fee
Apr
23
Online conference organized by Conf42.
This is a virtual event
This event requires an entrance fee
Apr
23
In-person conference organized by NDC.
Location: Sydney, AU
This event requires an entrance fee
Apr
24
In-person conference organized by Google.
Location: Las Vegas, NV, USA
This event requires an entrance fee
Apr
28
In-person conference organized by Devopsdays.
Location: Copenhagen, DK
This event requires an entrance fee
Kubernetes call for papers
27
days
Kubernetes Community Days Lima 2026
The Call For Paper is open until 19 May 2026 at UTC. More info →
Location: Lima, PE
In-person conference organized by KCD Lima, Perú.
The conference starts on the 18 July 2026.
- Apply here
11
days
The Call For Paper is open until 3 May 2026 at UTC. More info →
Location: Shanghai, CN
In-person conference organized by CNCF.
The conference starts on the 9 September 2026.
- Apply here
40
days
The Call For Paper is open until 1 June 2026 at UTC. More info →
Location: Bergen, NO
In-person conference organized by CND Norway.
The conference starts on the 27 October 2026.
- Apply here
43
days
The Call For Paper is open until 4 June 2026 at UTC. More info →
Location: Feira de Santana, BR
In-person conference organized by Devopsdays.
The conference starts on the 6 June 2026.
- Apply here
9
days
The Call For Paper is open until 1 May 2026 at UTC. More info →
Location: New York, NY, USA
In-person conference organized by SREday.
The conference starts on the 2 June 2026.
- Apply here
43
days
The Call For Paper is open until 4 June 2026 at UTC. More info →
Location: Curitiba, BR
In-person conference organized by Devopsdays.
The conference starts on the 22 August 2026.
- Apply here
11
days
The Call For Paper is open until 3 May 2026 at UTC. More info →
Location: Berlin, DE
In-person conference organized by Devopsdays.
The conference starts on the 29 September 2026.
- Apply here
40
days
The Call For Paper is open until 1 June 2026 at UTC. More info →
Location: Belgrade, RS
In-person conference organized by heapspace.
The conference starts on the 6 November 2026.
- Apply here
25
days
The Call For Paper is open until 17 May 2026 at UTC. More info →
Location: San Jose, CA, USA
In-person conference organized by TechEx Events.
The conference starts on the 19 May 2026.
- Apply here
Until next time!
— Gulcan