[–] 10 points 3 weeks ago

That's fair. I think they're are definitely systems where micro services are the best choice, but also that they are way over used or implemented too early. Too often it seems service boundaries follow the organization and not the domain model, because they only exist so multiple teams don't have to talk to each other as often. I guess I'm saying there are more incorrect ways of doing micro services than correct ways.

The worst I've seen though have been a single developer over engineering a system because they thought it was cool, but we replaced it with a simple batch script.

  • source
  • parent
  • context
  • [–] 11 points 3 weeks ago

    Completely unrealistic. The monolith is not retiring, it's just getting stuffed into a docker image and saying "Hello, fellow micro services!", while achieving a fraction of the old uptime because you're also trying to run Oracle or MSSQL on AWS using EBS as the underlying storage, and none of those components were designed for that use case.

  • source
  • [–] 25 points 3 weeks ago (1 child)

    I’ve worked on projects where microservices are a godsend. I’ve also worked on a 12-endpoint REST API atop a SQL database and an Elasticsearch instance serving 500 internal users that somebody had engineered into 8 microservices because management thought microservices were the future. Real mixed bag.

  • source
  • parent
  • context
  • [–] 2 points 3 weeks ago

    It's going to be a mess regardless. The monolith is just tangled at a different layer. The only way to avoid a tangled mess is to incentivize proper engineering, and everyone is incentivizing time to market and delivery of barely working features, which are the antithesis of proper engineering.

  • source
  • parent
  • context
  • [–] 35 points 3 weeks ago (5 children)

    Maybe an unpopular opinion, but I think if you're not sure if you need microservices, then you don't need microservices. I've had to clean up too many overengineered systems to be convinced otherwise.

  • source
  • [–] 5 points 3 weeks ago

    People love hating on stuff.

    I don't hate Matrix but I'm not a fan. My only experience was on mid-size chatrooms where that was enough to notice the usual e2e UX issues just amplified by the number of users. I also remember reading articles about security issues. Both of those things make me reluctant to use it.

    At the same time, I don't have a serious alternative to propose for multi-chatroom e2e messaging software (I haven't been looking though).

  • source
  • parent
  • context
  • [–] 6 points 3 weeks ago (1 child)

    Not to mention the stack traces. Gotta scroll through 100 abstraction layers just to find where your own code even starts... And that's if your code was the issue.

    Most days I actually like programming in Java, but these dynamic injection libraries are god awful to debug. I'm convinced that there is a growing number of vibe coders who have no idea how proper web services actually work.

  • source
  • parent
  • context
  • [–] 13 points 3 weeks ago (2 children)

    Nothing like debugging a Java Spring Boot application where every dependency is injected at runtime and nothing makes sense anymore and each class has hundreds of annotations, each getting progressively more psychotic.

  • source
  • [–] 3 points 3 weeks ago

    The requirements are pretty standard for what people want in a messaging protocol. None of those listed actually support those requirements so no one I know uses them. I don't know why people keep bringing those three up.

  • source
  • view more: ‹ prev next ›