Ask HN: Why do message queue-based architectures seem less popular now?

Re:Ask HN: Why do message queue-based architectures seem les... by xnickb on Wednesday June 19, 2024 @02:50AM (#40725549)

But my infrastructure is code! Can't you see how it's all in git?

[ Parent | on HN ]

Re:Ask HN: Why do message queue-based architectures seem les... by benterix on Wednesday June 19, 2024 @03:52AM (#40725916)

What's wrong with having your infrastructure as code and storing it in Git?

[ Parent | on HN ]

Re:Ask HN: Why do message queue-based architectures seem les... by xnickb on Wednesday June 19, 2024 @04:57AM (#40726301)

YAML isn't code. Same as your reply, YAML has very little awareness of context.

[ Parent | on HN ]

Re:Ask HN: Why do message queue-based architectures seem les... by benterix on Wednesday June 19, 2024 @06:41AM (#40726926)

In computer science, the word "code" has a very specific meaning, and markup languages definitely fit this definition.

[ Parent | on HN ]

Re:Ask HN: Why do message queue-based architectures seem les... by sofixa on Wednesday June 19, 2024 @04:58AM (#40726308)

Nothing. Even if it's objectively terrible (thousands of lines of templated YAML, or thousands of lines of spaghetti bash), being in Git as code is still better. At least you know what it is, how it evolved, and can start adding linting/tests.

[ Parent | on HN ]

Re:Ask HN: Why do message queue-based architectures seem les... by benterix on Wednesday June 19, 2024 @06:43AM (#40726934)

I manage large IaC repos and it's mostly HCL, well-structured and easy to work with. Where we have Kubernetes manifests, they're usually split into smaller files and don't cause any trouble as normally we usually don't deploy manifests directly.

[ Parent | on HN ]