619 posts tagged with "Cost Optimization"
Cloud vs owned-hardware cost breakdowns and bill analysis
Coolify vs Dokploy After a Real Migration: What the Same Laravel App on Two Identical Hetzner Boxes Reveals
The same Laravel app plus Postgres, deployed to Coolify and Dokploy on two identical Hetzner CPX21 boxes: install times, Railpack-vs-Nixpacks build speeds, and the one-click-vs-CLI restore gap that picked the winner — plus where the result confirms our second-machine showdown.
The Seat Is Dead: What Cost Per Accepted Change Means for Metering Agent-Driven Deploys
Coding-agent prices collapsed from seats to per-outcome cost in 2026. This post defines cost per accepted change, prices one agent deploy across three retry scenarios, and gives deploy platforms a four-rule metering model for MCP-driven deploy and rollback calls.
Leaving Heroku With Django? Inventory First, Price Table Second
A Django-on-Heroku migration inventory — buildpack pins, release-phase migrations, collectstatic, Celery topology, and Postgres extensions — that decides whether your exit to Render, Railway, Fly.io, or owned hardware is a weekend or a quarter, before any price table can tell you.
Fly.io's Edge Premium, Priced: A Single-Region Always-On App vs Render, Railway, and a $7 Hetzner Box
Fly.io's $2.19 headline assembles to $19 for an always-on single-region stack — priced line by line against Render (~$34), Railway (~$57), and a $7 Hetzner box, with an egress sensitivity table and the honest carve-out where the edge earns its keep.
Fly.io's Logs Went Dark for 4.5 Hours While Apps Stayed Up: What Owning Your Log Pipeline Actually Costs
Fly.io's historical logs lagged 4 hours 24 minutes on October 13 — its second log-ingestion incident in thirteen days — while apps kept serving. This teardown maps which half of the pipeline failed, prices 10 GB/month of logs across Datadog, Better Stack, Axiom, and a €4.35 self-hosted VictoriaLogs box, and shows how owning the log path turns the next backed-up queue into your own alertable metric.
Fly.io's 37-Region Edge vs the Two Regions You Actually Serve: Pricing the $3.88 Machine Against the $7 Heroku Dyno
Fly.io's $3.88 shared-cpu machine undercuts Heroku's $7 dyno per process — but the assembled two-region stack costs $30 vs $23 once Postgres enters, against ~$5–10 on flat Hetzner hardware, plus the edge ops overhead neither unit price shows.
From Azure to Hetzner: What a 2026 Switchers' Guide Gets Right About the Hyperscaler Exit Wave
A 2026 Azure-to-Hetzner switchers' guide promises 70–90% savings; this teardown verifies the egress math and unbundling map, prices the re-assumed ops labor, and distills a PaaS migrator's runbook.
Hetzner's 2026 Price Hikes and the 20TB-to-0.5TB Traffic Trap: Recomputing Self-Hosted Fleet Unit Economics
Hetzner raised cloud prices twice in 75 days in 2026 — the CCX13 jumped 169% to €42.99 while the ARM CAX11 held at €5.99 — and traffic allowances collapse from 20TB in the EU to 1TB in the US and 0.5TB in Singapore with €7.40/TB overage. This teardown recomputes per-vCPU, per-GB, and per-TB unit costs across the compute ladder, Server Auction metal, and all three regions, plus five rules for where just-add-another-box still works.
Hetzner Gives Away the Model Calls: What Free Inference Is Actually Worth to an Agent Builder
Hetzner's free experimental Inference API serves Qwen open-weight models at zero per-token cost — a worked comparison shows it replaces about $30 a month of hosted inference for a typical agent workload, within rate limits and no-SLA terms that rule out production use.