| |
| 🔗 Stories, Tutorials & Articles |
| |
|
| |
| v1.35: New level of efficiency with in-place Pod restart |
| |
| |
Kubernetes 1.35, as you may know, introduced in-place Pod restarts (alpha). It's a real reset: all containers, init and sidecars included - without killing the Pod or kicking off a reschedule. Think restart without the cloud drama.
Big win for workloads with heavy inter-container dependencies or massive AI/ML stacks. No more tossing the whole Pod just to clean house. It leans on extended Container Restart Rules and adds a new Pod condition for tracking. |
|
| |
|
| |
|
| |
| v1.35: Watch Based Route Reconciliation in the Cloud Controller Manager |
| |
| |
| Kubernetes v1.35 sneaks in an alpha feature gate that flips the CCM route controller from "check every X minutes" to "watch and react." It now uses informers to trigger syncs when nodes change - plus a light periodic check every 12–24 hours. |
|
| |
|
| |
|
| |
| 1.35: Enhanced Debugging with Versioned z-pages APIs |
| |
| |
Kubernetes 1.35 makes a quiet-but-crucial upgrade: z-pages debugging endpoints now return structured, machine-readable JSON. That means tools- not just tired humans - can parse control plane state directly.
The responses are versioned, backward-compatible, and tucked behind feature flags for now. |
|
| |
|
| |
|
| |
| From Deterministic to Agentic: Creating Durable AI Workflows with Dapr |
| |
| |
| Dapr dropped Durable Agents - a mashup of classic workflows and LLM-driven agents that can actually get things done and survive rough edges. They track reasoning steps, tool calls, and chat states like a champ. If things crash, no problem: Dapr Workflows and Diagrid Catalyst bring it all back. |
|
| |
|
| |
|
| |
| Implementing assurance pipeline for Amazon EKS Platform |
| |
| |
AWS released a full-stack CI/CD validation pipeline for Amazon EKS. It pulls in six layers of testing, Terraform, Helm, Locust load testing, and even AWS Fault Injection for pushing resilience to the edge.
The goal: bake policy checks, functional tests, and brutal load tests right into pre-deployment. Fewer surprises in prod. Less grunt work for ops. |
|
| |
|
| |
|
| |
| Streamline your containerized CI/CD with GitLab Runners and Amazon EKS Auto Mode |
| |
| |
| GitLab Runners now work with Amazon EKS Auto Mode. That means hands-off infra, smarter scaling, and built-in AWS security. Runners spin up on EC2 Spot Instances, so teams can cut CI/CD compute costs by as much as 90% - without hacking together flaky pipelines. |
|
| |
|
| |
|
| |
| Kubernetes GPU Management Just Got a Major Upgrade |
| |
| |
Kubernetes 1.34 dropped Dynamic Resource Allocation (DRA) - think persistent volumes, but for GPUs and custom hardware. Vendors can now plug in drivers and schedulers for their devices, and workloads can pick exactly what they need.
Coming in 1.35: a new workload abstraction that speaks the language of multinode jobs. It adds topological awareness and atomic scheduling, making big, messy AI workloads play nice with the cluster. |
|
| |
|
| |
|
| |
| BadPods Series: Everything Allowed on AWS EKS |
| |
| |
A security researcher ran a full-blown container escape on EKS using BadPods - a tool that spins up dangerously overprivileged pods. The pod broke out of its container, poked around the host node, moved laterally, and swiped AWS IAM creds.
All of it slipped past EKS’s default Pod Security Admission (PSA) policies. Why? Because those defaults still let pods declare risky stuff like hostPID, hostNetwork, privileged, and hostPath volumes. Basically, a welcome mat for escalation. |
|
| |
|
| |
👉 Got something to share? Create your FAUN Page and start publishing your blog posts, tools, and updates. Grow your audience, and get discovered by the developer community. |
