1732 posts tagged with "Self-Hosting"
Running your own PaaS and infrastructure on machines you own
Railway Patches Your Postgres CVEs and Ships MySQL and Redis HA: What That Managed-Data Moat Really Costs a Self-Hoster
Railway's August 2026 data-layer bundle — automatic Postgres CVE patching, MySQL and Redis HA, and CLI Postgres management — itemized line by line: what each job costs to replicate on owned hardware, what a typical app pays on the meter versus a flat Hetzner box, and why bex treats managed databases as a non-goal.
Render's $25 Flat Team Plan vs Railway's $20 Seats vs Fly.io's $0.02 Egress: Pricing the Same 5-Person Workload in 2026
We price one 5-person team workload — web service, Postgres, five preview envs, 500 GB egress — on Railway, Render, and Fly.io using July 2026 third-party benchmarks, then run the same workload as owned Hetzner capacity to show when the benchmark winner or the exit is the right call.
One Team Membership Was Enough: How Rancher CVE-2026-41053 Handed Every GitHub Org Member Every Team's RBAC Grants
Rancher CVE-2026-41053 (CVSS 8.8) let any user with one GitHub team membership inherit every team's Rancher permissions. Affected versions, the broken expansion loop, and the remediation checklist.
Render Will Pay You Up to $10K to Leave Railway: What Migration Credits Actually Cover
Render offers up to $10,000 in migration credits plus a migrate-from-Railway guide. A worked burn-down shows the credits buy years of compute but zero hours of engineering — plus the flat-hardware math that outlasts any credit.
Self-Hosted vLLM on Kubernetes: What Owning Your Inference Layer Actually Costs
CNCF's July 2026 vLLM walkthrough builds self-hosted inference in three Kubernetes resources. A worked breakeven — 10 to 15M tokens a day against Sonnet 5 on a €1,199/month GPU box — plus the four production gaps the lab skips and the five cases where owning the layer wins.
No SSH? No Problem: Talos Linux's talosctl debug Is a Privileged Escape Hatch That Cleans Up After Itself
Talos v1.13's talosctl debug runs a privileged container in the host's own namespaces — with registry-pull and offline tarball-push image paths — and deletes it on exit. We walk through a real pwru diagnosis and the audit rules for using it safely.
Boring Infrastructure on Purpose: What Uncloud's Control-Plane-Free Clustering Keeps — and What It Quietly Gives Up
Uncloud clusters Docker hosts across cloud VMs and bare metal with no control plane and no quorum — 5,200+ stars and a university moving 300+ sites onto it. A concrete keeps-versus-gives-up trade-off table, a dead-machine walkthrough, and why the verdict flips once AI agents operate the fleet.
Woodpecker CI's 130MB Control Plane vs a 7GB GitHub Runner: What Your PaaS Build Step Really Costs on a €3.79 Hetzner Box
A Woodpecker server and agent idle at about 130MB of RAM while GitHub's standard runner is a 2-core, 7GB VM. We budget both stacks megabyte-by-megabyte on a €3.79 Hetzner CX22 and work out when owning your build fleet beats paying per-minute Actions overage.
Rent the Sandbox or Own the Fleet? What a 26-Provider Survey Says About Agent Isolation
A 2026 survey lines up 26 agent sandboxes across startup time, isolation, pricing, and SDKs. The slot price has converged near $0.08/hour, snapshots are the only free lunch, and the math flips between renting and owning at 15-25% utilization.