510 posts tagged with "Kubernetes"
Container orchestration, Cluster API, and self-hosted control planes
From $7,000 on AWS to $780 on Hetzner: An Honest Accounting of an 89% Cost Cut
A team swapped a $7,000+ AWS Kubernetes bill for four €179 Hetzner servers — then listed what the $780 doesn't cover. A line-by-line audit of managed control planes, load balancing, backups, and on-call, plus the decision rule for when the 89% saving is real.
Buildpacks RFC 329 Would Make `web_worker` an Illegal Process Type — and That's the Point
Buildpacks RFC 329 would restrict process types to lowercase DNS labels, outlawing underscores, periods, and uppercase letters. Here is the exact rule, why the rejected 253-character hostname alternative lost, and what the migration costs buildpack authors and platforms.
cert-manager vs Gardener vs certctl: What Tenant TLS Really Costs at 10, 100, and 1,000 Custom Domains
A head-to-head comparison of cert-manager, Gardener cert-management, and certctl for per-tenant custom-domain TLS, with the rate-limit math and renewal checklist for fleets from 10 to 1,000 domains.
Firecracker Is the Easy 5%: What Self-Hosting an E2B-Style Agent Sandbox Really Operates
E2B's enterprise pitch admits the microVM is the easy part. An eight-piece control-plane inventory plus a hosted-vs-self-hosted cost comparison shows what running agent sandboxes yourself actually operates.
Ephemeral Preview Environments Without a SaaS Bill: Namespace-per-PR on Owned Hardware
Full-stack PR previews cost a typical 8-developer team about $99 a month on Render's cheapest seats — or roughly $0 marginal on an already-paid Hetzner box. A line-by-line cost comparison plus the ArgoCD namespace-per-PR wiring and TTL teardown discipline that make previews a default instead of a line item.
Talos vs Flatcar in 2026: Picking the Immutable OS Under Your CAPH Fleet
Talos Linux ships Kubernetes in an 80 MB API-only image with no SSH; Flatcar pairs familiar Linux with automatic A/B updates proven on 20,000 Adobe nodes. A head-to-head on provisioning, updates, 3am debugging, and governance — with a verdict for small CAPH fleets on Hetzner.
Flatcar Brings Omaha-Style Auto-Updates Into CNCF Custody: What A-B Partition Rollouts With Automatic Reboots Mean for Unattended Node Updates Across a Hetzner Fleet
Flatcar Container Linux entered the CNCF incubator with Adobe running 20,000 nodes on it. What its Omaha-driven A-B auto-updates change for Cluster-API fleets on Hetzner — and the channel pins plus reboot coordination that make unattended node updates safe.
Gateway API v1.6: One Ingress Stack for HTTP, TCP, and UDP — No Second Load Balancer
Gateway API v1.6 graduates TCPRoute and UDPRoute to stable v1, so one Gateway can serve HTTP apps, TCP services, and QUIC/UDP endpoints. A concrete one-Gateway design with YAML, the TLS terminate-vs-passthrough table, L4 port and tenancy rules, and when to keep a second load balancer anyway.
Headlamp's Plugin Sprint Gave Kubeflow, Volcano, and Knative a GUI: Is It Ready to Be Your Fleet's Operator Console?
Headlamp shipped GUI plugins for Cluster API, Volcano, Knative, and Kubeflow in summer 2026. A plugin-by-plugin scorecard — versions, capabilities, honest gaps — and what it means for a self-hosted fleet running AI-agent workloads.