[debrice]

The best hands-on practice was when I had to build my own, nobody to tell me to cut corners and time to do things right. It's important to do it right to know how to do it wrong (most company will prevent you from doing things right because of time/budget constraints).

Ask yourself:

- What constitute a _good_ infrastructure?

- Something flexible/adaptable ?

- Something predictable ?

- Something resilient ?

- Something easy to monitor ?

- Something else ?

- How do you prioritize these, who gets to decide which one matters most...

Architecture is asking a lot of questions, inviting others who will benefit/use your work into the decision making process.

One thing that I learned that is still helpful to me today is try to push as many decisions as possible to later, keep the door open for futur choices. Every system architecture is part invention, part discovery. As you put together your foundation, new requirements/restrictions will appear and you'll be happy to have the freedom to make some of your decisions then.