The post you’re reading right now was published by a Python script running on a 4 vCPU Hetzner box in a data center I’ve never seen. That server has been up for twenty weeks straight. It runs the Postgres database, the content pipeline, the quality checks, and the WordPress site itself, and it does all of that for less than what I spend on coffee in a normal week. So when people ask me “Hetzner vs DigitalOcean,” I’m not answering from a benchmark table. I’m answering from my invoices, because I’ve paid both companies real money over the years and I still have opinions about both bills.<\/p>\n
Short version for the impatient: Hetzner for compute, DigitalOcean for hand-holding. If you want to know why, and where each one will bite you, read on.<\/p>\n
Let me put actual numbers on the table, because most VPS comparisons are weirdly shy about them. My Hetzner server has 4 vCPUs, 8GB of RAM, and an 80GB disk, of which I’m using 17GB. It runs Ubuntu 24.04. On it: Postgres, a Python content pipeline that drafts and publishes two blog posts a day, nginx, and the WordPress install you’re reading. Memory sits around 1.4GB used. The machine is, frankly, bored.<\/p>\n
That spec level is where Hetzner’s pricing gets hard to argue with. I’d encourage you to open Hetzner’s cloud pricing<\/a> and DigitalOcean’s pricing<\/a> side by side and do the math for your own workload, because numbers change and I’d rather you distrust me than quote a stale table. But the shape of the gap has held for years: for the same monthly spend, Hetzner has consistently handed me roughly twice the cores and RAM, and the included bandwidth is generous enough that I’ve never once thought about it. On DigitalOcean I have thought about bandwidth. Thinking about bandwidth is my least favorite hobby.<\/p>\n Here’s the thing the raw price comparison misses: DigitalOcean isn’t really selling you a VPS. It’s selling you the stuff around the VPS.<\/p>\n The DO community tutorials are still some of the best server documentation on the internet, and I say that as someone who no longer pays them. When I was learning this stuff a decade ago, half of what I knew about nginx and firewalls came from DO tutorials I followed on other providers’ machines, which in hindsight was a bit rude of me. Their managed Postgres means someone else owns your backups and failover. The app platform means a junior dev can deploy without ever learning what systemd is.<\/p>\n None of that is nothing. If you’re a team without an infrastructure person, the DO premium is cheaper than the outage where you discover your backup cron job had been silently failing for eight months.<\/p>\n Hetzner sells you a very good, very cheap computer, and then it wishes you luck. There’s no managed database. The console is functional and Germanic. Support is fine but lean. New accounts sometimes hit identity verification friction that feels heavy-handed until you realize how much abuse a cheap-compute provider attracts. You’re expected to know what you’re doing, and the price assumes you do.<\/p>\n The day-one experience is nearly identical. Both have solid CLIs and both take the same cloud-init files:<\/p>\n Ten seconds either way and you have a root shell. The divergence starts on day two. On DigitalOcean, when you need a database, a load balancer, or object storage, there’s a button. On Hetzner there’s a longer path: you install Postgres yourself, you configure the firewall yourself, and your “load balancer” might be you reading the docs at 11pm. I happen to enjoy that path, and my setup reflects it: my scheduled jobs run on systemd timers instead of cron<\/a>, and my TLS renews itself since I moved from nginx to Caddy on side projects<\/a>. But I earned that comfort through years of breaking things, some of them for clients.<\/p>\n Location matters more than people admit, too. DigitalOcean has data centers spread across the US, Europe, and Asia. Hetzner is strongest in Europe with a smaller US and Singapore footprint. I’m in Dubai and my audience is global, so European latency is a fine trade for me. If your users are in Ohio, measure before you commit.<\/p>\n I owe you the failure modes, since I’m clearly the cheap-compute partisan here.<\/p>\n First, everything is your fault. That’s the deal. When my disk fills up because a log rotation config was wrong, there’s no support plan that catches it. My pipeline logs its own alerts into Postgres precisely because nobody else is watching.<\/p>\n Second, backups are on you. Hetzner sells snapshot and backup options, and they’re cheap, but turning them on, testing a restore, and actually doing a restore drill is your job. An untested backup is a rumor, not a backup.<\/p>\n Third, if you ever want to hand the infrastructure to a non-technical client, the DO control panel is something I can walk a client through on a call. The Hetzner console plus a folder of my shell scripts is not. A couple of my client projects live on DigitalOcean for exactly this reason, and it’s the right call there, even at the higher price.<\/p>\n For your own projects, if you’re comfortable in a terminal: Hetzner, without much hesitation. The savings compound. The money I don’t spend on hosting this blog pays for the AI API calls that help produce it, which feels like a fair trade between robots.<\/p>\n For client work where the client inherits the infrastructure: DigitalOcean, or at least something managed. You’re not buying compute, you’re buying the ability to walk away without guilt.<\/p>\n For a team with no ops experience and a product to ship: DigitalOcean’s managed services, and revisit the bill when it starts to hurt. The migration from managed Postgres to self-hosted is much easier once you have revenue and much harder at 2am with none.<\/p>\n And whichever you pick, the boring parts matter more than the provider. Automated backups you’ve actually restored once, a firewall that closes everything but 22, 80, and 443, and unattended security updates will do more for you than any comparison post, including this one. This whole setup is part of how I run my one-person studio<\/a>, and the server bill is the least interesting line in the budget.<\/p>\n Spin up the cheapest box on whichever provider you’re curious about. Both bill hourly, so an evening of poking around costs less than a bus ticket. Deploy one real thing on it: a static site, an uptime checker, anything you’d mildly miss if it died. Then turn on backups, break the thing on purpose, and restore it. That last step is the whole lesson. Twenty weeks of uptime on my box isn’t luck, it’s the three times I practiced un-breaking it before it mattered.<\/p>\n","protected":false},"excerpt":{"rendered":" The post you’re reading was published by a Hetzner box that’s been up for twenty weeks. My honest Hetzner vs DigitalOcean math after paying both providers.<\/p>\n","protected":false},"author":2,"featured_media":642,"comment_status":"","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"rank_math_title":"","rank_math_description":"The post you're reading was published by a Hetzner box that's been up for twenty weeks. My honest Hetzner vs DigitalOcean math after paying both providers.","rank_math_focus_keyword":"hetzner vs digitalocean","rank_math_canonical_url":"","rank_math_robots":"","footnotes":""},"categories":[302],"tags":[80,702,701,8,703],"class_list":["post-643","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-devops","tag-devops","tag-digitalocean","tag-hetzner","tag-self-hosting","tag-vps"],"_links":{"self":[{"href":"https:\/\/abrarqasim.com\/blog\/wp-json\/wp\/v2\/posts\/643","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/abrarqasim.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/abrarqasim.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/abrarqasim.com\/blog\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/abrarqasim.com\/blog\/wp-json\/wp\/v2\/comments?post=643"}],"version-history":[{"count":0,"href":"https:\/\/abrarqasim.com\/blog\/wp-json\/wp\/v2\/posts\/643\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/abrarqasim.com\/blog\/wp-json\/wp\/v2\/media\/642"}],"wp:attachment":[{"href":"https:\/\/abrarqasim.com\/blog\/wp-json\/wp\/v2\/media?parent=643"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/abrarqasim.com\/blog\/wp-json\/wp\/v2\/categories?post=643"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/abrarqasim.com\/blog\/wp-json\/wp\/v2\/tags?post=643"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}What DigitalOcean is actually selling<\/h2>\n
Provisioning is a coin flip, the rest isn’t<\/h2>\n
# Hetzner\nhcloud server create --type cx32 --image ubuntu-24.04 \\\n --ssh-key mykey --user-data-from-file cloud-init.yml blog-server\n\n# DigitalOcean\ndoctl compute droplet create blog-server --size s-4vcpu-8gb \\\n --image ubuntu-24-04-x64 --ssh-keys $KEY_ID --user-data-file cloud-init.yml\n<\/code><\/pre>\nWhere Hetzner will bite you<\/h2>\n
My actual recommendation<\/h2>\n
Try this week<\/h2>\n