Why your apps keep running when we do not

Este documento está disponível apenas em inglês. Traduzi-lo mudaria o que ele promete, e uma tradução que divergisse do original seria justamente a que você leu. O restante do DevOps Buddy está disponível no seu idioma.

Most platforms that deploy your app also run it. That is convenient right up until the platform has a bad afternoon, and then your site has one too.

DevOps Buddy is built the other way round. Your apps run on your virtual machines, in your cloud account, behind a reverse proxy on that same machine. We hold the buttons, not the traffic.

What that means in practice

If our control plane is unreachable, three things are true:

  1. Every site you have deployed keeps serving. The traffic never passed through us.
  2. TLS certificates keep renewing. Let's Encrypt talks to your server, not to us.
  3. You cannot make changes — no deploys, no new servers — until we are back.

That third one is the honest cost, and it is the one we accept. A deploy that has to wait an hour is an inconvenience. A site that is down for an hour is a phone call from a customer.

The agent only speaks outwards

The piece of us that runs on your machine is a single Go binary. It polls us over HTTPS and asks whether there is anything to do. It does not listen on a port, and there is no SSH key of ours on your server.

That means you can put your machine behind whatever firewall you like — Buddy does not need a hole in it.

Leaving

On a container server, your apps are ordinary docker compose projects under /var/lib/buddy/apps. On a classic server, they are ordinary nginx sites under /var/www. Stop paying us, remove the agent, and everything keeps running exactly as it is.

That is the part we would want to read before trusting a platform, so it is the part we built first.


Rodar isso na sua própria cloud leva uns dez minutos. Começar de graça