1512 posts tagged with "PaaS"
Platform-as-a-Service tooling, economics, and self-hosted alternatives
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.
Render Never Ran Your Heroku Buildpacks — Here's What Rebuilding That Compatibility Costs With Paketo vs Nixpacks
Render never executed your Heroku buildpacks — it absorbed them behind a generated Dockerfile. Here is what rebuilding that compatibility on machines you own actually costs, with Paketo and Nixpacks priced head-to-head.
Heroku Dyno to Hetzner CX in 2026: Real Bill Before and After With Egress Math
The same Node.js and Postgres workload costs $128 a month on Heroku Standard dynos versus about $10 on a Hetzner CX box. A line-by-line 2026 recompute with egress math from 100 GB to 20 TB, plus the honest case for staying on Heroku.
Heroku's Dead, Three Bills Remain: Railway vs Render vs Fly.io vs a Flat-Rate Hetzner Box
Techsy's July 2026 comparison priced the same app on Railway, Render, and Fly.io. We add the missing fourth column — flat-rate Hetzner machines — and show where each platform's billing shape hides the cost that owning the box eliminates.
k0rdent Turns One: What a Year of Shipping a Cluster API Fleet Orchestrator Proves About Not Building Your Platform From Scratch
k0rdent went from v0.1 to v1.7 in its first year, and RBC Capital Markets now runs 50+ clusters on it. What the Cluster API-based orchestrator's lifecycle, state, and observability layers prove — and which parts a small self-hosted fleet should borrow, skip, or work around.
Kubernetes 1.36 Teaches Controllers to Distrust Their Own Cache: Why a Skipped Sync Beats a Wrong Reconcile
Kubernetes 1.36 makes DaemonSet, StatefulSet, ReplicaSet, and Job controllers skip any sync where the informer cache lags their own last write. The mechanism, the two metrics that expose it, and why a lean self-hosted fleet feels every wrong reconcile as an invoice.
Lint the Dockerfile Before You Build It: What hadolint, dockle, and Docker Scout Catch That a Green Build Never Will
A green docker build waves through unpinned base tags, root users, baked-in secrets, and known CVEs. How hadolint at PR time, dockle at build time, and Docker Scout at deploy time each catch a failure class the others cannot see — with gate configs you can copy.
MCP Tasks and Multi-Round Trips: The Durable Deploy Contract an Agent-Operated PaaS Needs
MCP's July 2026 spec adds asynchronous Tasks and Multi-Round Trip Requests. Together they form a durable deploy state machine: task handles instead of blocked calls, an approval boundary before production promotion, safe retry, cooperative cancel, and rollback as a first-class task.