You're going to have outages in production. They're inevitable. The question is how to best minimize outages, both their frequency and duration. Common wi...

cross-posted from: https://lemmy.world/post/21023181

Sharing some lessons I learned from 10 years/millions of users in production. I’ll be in the comments if anyone has any questions!

I hope some of the lessons in this series help people learn to adopt Linux directly into their stack as a simple tool that can be managed easily on a server.

  • @[email protected]
    8•1 month ago

    The amount of complication in “modern” pipelines is insane. If you need a dozen pieces of software just to go live, you’re making things difficult.

      • @[email protected]
        2•1 month ago

        I guess. I mostly think it’s ignorance. Relying on a specific architecture or platform, etc. is super dangerous—and costly.

    • Possibly linuxEnglish
      2•1 month ago

      The problem is all the solutions. Just make some automation that is predicable and automated. You can use automation to build the container and then run tests on it. If the tests pass you can do a gradual roll out. All of this can be simple automations that are either Ansible playbooks or bash scripts.

  • Possibly linuxEnglish
    4•1 month ago

    I think Kubernetes has its value. It can be wildly complex but it also can be reasonable and clean. For my Homelab I like k3s as it is lightweight and gets the job done. I can setup automatic rollbacks and have all the features of Kubernetes without the complexity of the commercial providers.

    I’m sure there are other Kubernetes distributions that do the same.

  • @[email protected]
    1•1 month ago

    There is very little content here.

    Could have boiled it down to modern CI/CD is over complicated, I’m going to talk about it. Then why not start talking about it with point 1 rather than just saying you will.

    I have no idea whether the content will be meaningful or not.

    More killer, less filler next time.