Blog
Insights, analysis, and updates from the AI agent economy. Browse by tag · Browse the archive.
Managed Control Plane vs Free Panel vs Hosted PaaS: What Deploynix's $12 Middle Ground Really Costs
A three-service stack with Postgres and Redis runs about $6/month on a self-run Coolify panel, $18 with Deploynix's $12 managed control plane, and $38 on Render — until you price the pager hours. A worked cost-plus-failure-mode comparison across Coolify, Dokploy, CapRover, Deploynix, Render, and Railway.
llm-d Is Now a CNCF Sandbox Project: What Kubernetes-Native Distributed Inference Means for Self-Hosting Your AI Agents' Brains
llm-d brings distributed LLM inference — prefix-cache-aware routing and prefill/decode disaggregation — to Kubernetes as a CNCF Sandbox project. What it means for self-hosting AI-agent inference, what it costs versus hosted APIs, and how early the project still is.
45-Day Certificates Are Here: What Let's Encrypt's Shrinking Validity Window Demands From Your Renewal Automation
Let's Encrypt now issues 45-day certificates, and 64-day defaults land in February 2027. Here is the full timeline, the renewal-window math, and a checklist to harden ACME automation before the manual-fix window disappears.
LangGraph Platform Is GA: Do Your Agents Need a Per-Node Runtime, or Just a Container?
LangGraph Platform's GA brings per-node agent billing to production. We work out the cost math at typical volumes, separate the free primitives from the metered ones, and give a rule for when to pay versus shipping a container.
No Terraform Required: KubeOne 1.13's One-Manifest Hetzner Control Planes vs Cluster API
KubeOne 1.13 provisions Hetzner control-plane VMs and the API load balancer straight from one manifest — no Terraform state, no output plumbing. Here is what the beta removes, what it still lacks next to Cluster API, and which operator each tool fits.
kube-hetzner Just Made Leap Micro the Default Node OS: What the Switch Off MicroOS Means for Your Hetzner Cluster
kube-hetzner's v3 line boots Leap Micro on every new nodepool while existing MicroOS nodes stay put. What the switch means, why fixed-point releases beat rolling ones for production nodes, and how Leap Micro compares to Talos.
Ingress-NGINX Is Retired — Your Gateway API Migration Can No Longer Wait
Ingress-nginx retired in March 2026 with no more security patches. A concrete migration map from Ingress annotations to Gateway, HTTPRoute, and GatewayClass — what ingress2gateway converts, what it skips, and where Gateway API v1.5 is genuinely better.
HPA Scales to Zero in Kubernetes v1.37: The End of the Always-On Replica (and the One Piece You Still Have to Build)
Kubernetes 1.37 graduates HPA scale-to-zero to beta, so idle tenant services can drop their last replica with no add-on — but HTTP workloads still need an activator. The worked cost math and a git-push PaaS adoption playbook.
Hetzner's Warsaw PoP and the GEX45: What Changes — and Doesn't — for a CAPH Fleet
Hetzner's Warsaw PoP improves Eastern-European peering but adds no region to target, while the €214/month GEX45 puts Blackwell GPUs on dedicated boxes CAPH can only adopt through Robot inventory. The numbers and the operator playbook for a Hetzner-pinned fleet.
Subscribe
New posts land in your reader as soon as they publish. Pick a format — all three carry the same posts.
Current feeds keep roughly two days of posts so daily polling does not miss a burst. Older entries stay reachable from the feed's next-page link in readers that follow it, or from the blog archive.
Following one topic instead? Browse tags