Pinterest's MCP Servers Save 7,000 Engineering Hours a Month: What 66,000 Invocations Teach About Agent-Operated Infrastructure ROI
Pinterest's production MCP ecosystem handles 66,000 monthly tool calls across 844 engineers, saving an estimated 7,000 hours a month. A teardown of where those savings came from, how solid the math is, and which operations a smaller platform should expose to agents first.
Dokploy Killed Its Global Build Queue After Two Years: What Per-Server Build Slots Reveal About Scheduling Outgrowing One Box
Dokploy ran every build on every server through one global queue for two years — then v0.29.11 gave each server its own. The wait-time math behind the fix, where per-box caps still strand capacity, and what a fleet build pool owes default-1.
Gateway API v1.5 Moves Six Features to Stable: Which Ones Your Git-Push PaaS Should Adopt First
Gateway API v1.5 graduates ListenerSet, TLSRoute, and four more features to the stable channel. A verdict table and migration checklist for platforms weighing the typed routing API against Ingress annotations.
Hardening an Infrastructure MCP Server: TLS, Scoped Credentials, and Rate Limits Before an Agent Touches Production
An MCP server that can redeploy production is a deploy pipeline with natural-language input. A seven-control hardening checklist — TLS, per-tool scopes, short-lived credentials, schema validation, rate limits, and audit logging — mapped to the three attacks that actually happen.
kops Is Tracking Kubernetes 1.37 in 2026 — and Still Proves Why Cluster API Won the Heterogeneous Fleet
kops shipped 1.36 and already tracks Kubernetes 1.37 — yet its own release notes show why declarative Cluster API won fleets spanning clouds and bare metal, and which upgrade disciplines a self-hosted platform should steal from it.
Deploy Agents That Wait: Building a Stateless Rollout-and-Approval Agent on MCP Tasks and Multi Round-Trip Requests
MCP's July 2026 release made the protocol stateless and gave agent builders Tasks and Multi Round-Trip Requests. Here is the concrete design for a deploy agent that returns a task handle, polls rollout status against Kubernetes Deployment conditions, and stops for an expiring human approval before promoting to production.
Railpack Is Railway's Default Builder: What the Nixpacks-to-BuildKit Rewrite Teaches Self-Hosted PaaS Builders
Railway replaced Nixpacks with Railpack, a Go rewrite on raw BuildKit, cutting base images 38-77%. A look at what broke, what the numbers prove, and what CNB-based self-hosters should borrow — plus where the Dockerfile escape hatch still wins.
Railway's May 2026 GCP Blackout: What an 8-Hour Outage Reveals About a Control Plane That Never Actually Left Google Cloud
Google suspended Railway's production account and healthy bare-metal workloads went dark for eight hours — because the control plane routing them never left GCP. A timeline of the cascade and a seven-question audit for your own platform.
Railway Retired Nixpacks for Railpack: What Changed in the Zero-Config Build Path
Railway replaced Nixpacks with Railpack after 14 million builds exposed the cost of approximate versions and monolithic image layers. A before-and-after of the auto-detect build path, why Cloud Native Buildpacks lost twice, and what platform builders should steal.