Hetzner Hostage
In this technical retrospective (2014), @tluyben2 shares practical architectural lessons learned from running, scaling, or migrating production software under lean constraints.
Core Principles for Independent Builders:
Defaulting to simpler architectures (such as monolithic frameworks, SQLite, or managed VPS) prevents early engineering bankruptcy.
Evaluating true compute and egress costs reveals that high cloud bills can often be slashed by 80%+ through judicious re-platforming.
Code you write at 2 AM as a solo engineer must be easy to debug six months later without complex microservice dependencies.
Community Debate Themes:
HN engineers debated the tradeoffs between cloud convenience and raw server economics, database indexing strategies, and resisting premature architectural complexity.
CATEGORY Frugal Engineering →
HISTORICAL ERA Cloud & Mobile Era (2014–2019)
COMMUNITY SCORE 50 points
TOTAL DISCUSSIONS 22 comments
I'm trying to think of the economic way to run a robust small scale service. I think you need a way to rapidly spin up a replica service on an entirely different provider.
I'm picturing a setup where you run on a cheap Hetzner host or similar with the DB synced to a slave replica on EC2 or other cloud provider and a build system so that you can spin up a whole replacement infrastructure on EC2 if there is a severe outage or failure in commercial relationship and switchover by changing the DNS settings.
One of the big political blogs in New Zealand recently had a similar problem with Linode.
They got DOSed so Linode shut them down, then they found a new provider but had problems moving to them since they couldn't get into the linode server to copy the data.
http://www.whaleoil.co.nz/2014/01/ferals-faults-fixes-co-pilots-report/ ↗
You have two ISPs for any business-critical service that matters, and you switch between them either using BGP in your router or using DNS if you have not implemented BGP.
With a single provider, it's a gamble, although the expectation for any reasonable provider is that they just let you saturate your link speed with no questions asked, and block the offending traffic in their router if you can identify it for them.
No story here. No budget provider I've looked at says they will help you withstand a DDoS attack. [Thanks to the commenter who says OVH does, I will look into them closely.]
I did this due diligence for a 10-user app. These guys have no excuse for not planning for a DDoS with a serious business.
You should have a look at OVH, they include DDoS protection by default and their prices are the best available on the market:
http://www.ovh.com/us/dedicated-servers/enterprise/ ↗