Zeropod Enables Stateful Kubernetes Scale-to-Zero via Container Checkpointing 📌 Zeropod revolutionizes Kubernetes by letting stateful apps scale to zero without losing a single byte of memory-freezing them mid-session for instant restarts when traffic returns. No more painful warm-ups or resource waste; pods wake up in milliseconds, preserving everything from files to network connections. A game-changer for stateful workloads that demand reliability and efficiency. 🔗 Read more: https://lnkd.in/dGtvtjAV #Zeropod #Kubernetes #Containerd #Checkpointing #Stateful

☸️ Kubernetes won. But at what price? Ten years after its first commit, Kubernetes has become the default foundation of modern infrastructure — from startups to global enterprises 🌍 But success has its shadows: 🧩 Growing complexity 👩💻 Talent shortages 💸 Rising cloud costs 🏗️ Organizational pressure many teams underestimate In his thought-provoking article, @maxkoerbaecher explores why Kubernetes is just the starting point — not the final destination. He dives into: 🚀 The hidden organizational shadow behind platform complexity ⚙️ The socio-technical limits of DevOps & tool overload 🔮 What comes next: WebAssembly, Platform Engineering & AI-driven operations ☁️ Why cloud strategies may need a radical rethink 📖 Read the full article: https://lnkd.in/dxVK8T6P #Kubernetes #CloudNative #PlatformEngineering #DevOpsCon #WebAssembly #CloudStrategy

Now, Localstack is no longer available, the free tier model, does not allow you to run tests on a CI/CD Pipeline. Depending on the complexity of your integrations tests you can try MotoServer. But I'll let an important note: If most of your tests are strong couple to AWS SDK, maybe you should make a refactor! https://lnkd.in/d73Jx3PK

I've been running both Coolify and Dokploy in my homelab for a few weeks. Both are genuinely impressive tools. But they share a hidden assumption that catches a lot of people out. They assume you have an empty server. Fresh VPS from scratch? Fantastic — Heroku-like simplicity, clean defaults, everything works in minutes. Existing server where you already run Traefik, manage your own Docker networks, and handle your own certs? Pain. These tools want to own ports 80 and 443, the network routing, the cert management. They don't integrate with what's already there — they replace it. This is why I keep coming back to a modular approach: separate tools for CI/CD, configuration management, and container management that each do one thing well and compose cleanly. None of them tries to own the machine. That said — starting fresh on a new server? Both are worth a serious look. The tradeoff is real, but so is the convenience. My full Coolify vs Dokploy comparison: https://lnkd.in/dssAWhe9 Which approach are you running — all-in-one PaaS or modular toolchain? #selfhosted #platformengineering #infrastructure #docker #devops

Recent days friends have started asking me how you manage Token costs and if i have several subscriptions :) My answer: Always "Edge inference " 😀 🤖 😀 Claude Code without paying out of pocket. Download Ollama https://ollama.com Install the Qwen model ollama run qwen3.5:9b Launch it from the terminal ollama launch claude --model qwen3.5:9b Done ! For bigger orchestration I use openclaw configured on Free tier AWS and Claude subscription to orchestrate and kimi for agent swarm

“Identify those savings and keep them as an ace up your sleeve while using that as leverage to get your roots in the system.” - Henrique Campos Amorim In this week’s episode, Taylor Houck sat down with Henrique Amorim, Business Analyst and FinOps Specialist at Crayon and Founder of FinOps Foundation in Denmark. Henrique shares why FinOps is more than just the “cost police” and how to use savings as leverage to show value. They talked about: - The difference between theoretical savings and truly actionable optimizations - The importance of building dashboards, processes, and visibility before scaling savings - How to communicate FinOps impact to executives - The role of community in accelerating your FinOps maturity - Why some “cost inefficiencies” are actually architectural features 🔗 Catch the full episode now with the links in the comments! FinOps In Action is brought to you by PointFive, which empowers teams to optimize cloud costs with advanced detection and remediation tools that drive action.

This week on FinOps in Action, I sat down with Henrique Campos Amorim to chat about why FinOps is not just the “cost police”. We also covered AI, the power of FinOps communities, and how to effectively show value to executives. Check out the episode - links in the comments!

Platform teams running Kubernetes spend a surprising amount of time on work that has nothing to do with their applications. Node provisioning, OS patching, GPU driver alignment, cluster component upgrades. At re:Invent 2025, Sai Vennam and Alex Kestner from AWS, joined by Daniel Levine from Capital One, spent an hour on how EKS Auto Mode changes that equation. I wrote it up so you don't have to sit through the full session. The article covers the shift in Kubernetes shared responsibility that Auto Mode brings, a live demo deploying a 20-billion-parameter LLM on a GPU spot instance with zero driver configuration, and Capital One's adoption story in a regulated enterprise. Dan's measure of success was silence in their Slack channels. 😊 🔗 https://lnkd.in/dR9384ZQ

The most expensive lines of code are the ones you haven't deployed yet. One Terraform line can burn $10,000 in a weekend. And you won't know until the bill arrives 30 days later. This is the "Reactive FinOps Trap." We spend hours cleaning up waste. We should be spending minutes preventing it. You need to catch costs in the Pull Request, not the Invoice. I built the FinOps IaC Checklist to stop the bleeding before it starts. It automatically flags the "Silent Killers": 1. The Zombie Instances: Flag missing auto-shutdown tags on Dev environments. 2. The Open Wallet: Detect missing budget alerts or guardrails. 3. The Oversized: Identify instance types that are 2x bigger than needed. 4. The Wild West: Prevent unrestricted auto-scaling groups (ASGs). Grab the full checklist by registering to our summit: https://lnkd.in/eEmwwnFK P.S. How often do you check cloud costs: Daily, Weekly, or only when Finance screams?

After some nice traveling back to shipping again! Just released Uncloud v0.17 🚀 Think multi-machine Docker Compose for production Rolling deployments now monitor crashes + health checks after new containers start. If a container keeps restarting or turns unhealthy, the deployment automatically rolls it back to prevent downtime. Enjoy safer deploys by default. And new docs on rolling deployments