Hacker News Reader

Comment by jimschley | Hacker News Reader

We went through this recently at Codeship when our provider, DNSimple, had an outage due to DDoS- https://blog.codeship.com/dnsimple-ddos-outage/ DNS is a service that often ends up as a single point of failure in infrastructures I've seen as it's non-trivial to implement redundancy. Having a repository/API approach to deploying DNS records saved us in this incident: http://blog.codeship.com/dnsimple-dns-history-continuous-dep...

View on HN

Nice to see you mention a repository - I love storing DNS details in git, and setup a site to push that on to Amazon's route53 infrastructure (https://dns-api.com/).

Having revision-control is wonderful for history-tracking.

This is a really cool idea. I'd like it even more if I could add "remotes" for different providers. DNSimple, Route53, etc. and push my changes to each one. I suppose the service would need to support a number of providers that themselves support APIs.

Would also be possible to use the dns_deploy tool one of my team mates wrote (https://github.com/codeship/dns_deploy/commits/master) or by extending Terraform (https://terraform.io/docs/providers/dnsimple/index.html)

Hey I wanted to thank you for showing this to me. I am using it on my personal domain and plan to use it on some customers as well. Really cool tool, thanks!

I did think of that, actually hosting the git repositories, and having a hook to make the updates on pushs.

But it seems like I can be significantly simpler and more likely to be reliable if I just get triggered by a webhook. I think I'd be willing to do it for a couple of people as a trial but I'd not expect so much interest.

Reply on news.ycombinator.com