Silas: I want to wash my car. The car wash is 50 meters away. Should I walk or drive?
Gemini:
….
That is a classic “efficiency vs. logic” dilemma.
Strictly speaking, you should drive. Here is the breakdown of why driving wins this specific round, despite the short distance:
...
* The “Post-Wash” Logic: If you walk there, you’ll eventually have to walk back, get the car, and drive it there anyway. You’re essentially suggesting a pre-wash stroll.
When should you walk?
…
3. You’ve decided the car is too dirty to be seen in public and you’re going to buy a tarp to cover your shame.
Seems pretty similar to https://github.com/radioactive-labs/chrono_forge which is what I found when I typed in "rails durable execution patterns" into Google. Have you seen this and if so, how do you think it compares?
I'm not trying to take a shot at the OP, but I keep seeing posts labeled "Production-Grade" that still look more like pet systems than cattle. I'm struggling to understand how something like this can be reproduced consistently across environments. How would you package this inside a Git repo? Can it be managed through GitOps? And if we're calling something production-grade, high availability should be a baseline requirement since it's table stakes for modern production applications.
What I'd really love is a middle ground between k8s and Docker Swarm that gives operators and developers what they need while still providing an escape hatch to k8s when required. k8s is immensely powerful but often feels like overkill for teams that just need simple orchestration, predictable deployments, and basic resiliency. On the other hand, Swarm is easy to use but doesn't offer the extensibility, ecosystem, or long-term viability that many organizations now expect. It feels like there's a missing layer in between: something lightweight enough to operate without a dedicated platform team, but structured enough to support best practices such as declarative config, GitOps workflows, and repeatable environments.
As I write this, I'm realizing that part of the issue is the increasing complexity of our services. Every team wants a clean, Unix-like architecture made up of small components that each do one job really well. Philosophically that sounds great, but in practice it leads to a huge amount of integration work. Each "small tool" comes with its own configuration, lifecycle, upgrade path, and operational concerns. When you stack enough of those together, the end result is a system that is actually more complex than the monoliths we moved away from. A simple deployment quickly becomes a tower of YAML, sidecars, controllers, and operators. So even when we're just trying to run a few services reliably, the cumulative complexity of the ecosystem pushes us toward heavyweight solutions like k8s, even if the problem doesn't truly require it.
I've been looking at this for gunzip files as well. There is a rust solution that looks interesting called https://docs.rs/indexed_deflate/latest/indexed_deflate/. My goals are to be able to index mysql dump files by tables boundaries.
While not specific to 12factor question. With any of these agents and solutions how is LLM Ops being handled? Also, what's the testing strategy and how do I make sure that I don't cause regression?
Nice, I'm working in the same space as you (not opensource, personal project). We landed on the same solution, encoding the commands inside Golang and distributing those via SSH.
I'm somewhat surprised not to see this more often. I'm guessing supporting multiple linux versions could get unwieldy, I focused on Ubuntu as my target.
Differences that I see.
* I modeled mine on-top of docker-plugins (these get installed during the bootstrapping process)
Your solution looks much simpler than mine. I started off modeling mine off fly.io CLI, which is much more verbose Go code. I'll likely continue to use mine, but for any future VPS I'll have to give this a try.
Do you have any sources on GitHub moving away from Rails? This is the first that I've heard and my googlefu has returned zero results around this. Just last year they had a blog post around Building GitHub with Ruby and Rails[0] so your remark caught my off guard.
I don't mind this style, it's encouraged by dry-monads with the do-notation. The big concern that I have with this code is it's missing a service level transaction.
This works pretty good, but we've ran into issues where the downloaded file isn't code signed and then this becomes a big issue. AFAIK, if this is a public CLI then homebrew will codesign for you, if you are distributing a private CLI then you need to solve the code signing issue yourself. Not many blog posts on how to handle this...
Those built-ins that you are talking about are not as well rounded as a dedicated framework. These built-ins operate at a different operating model than running additional background servers. This typically means that everyone in the org needs to be an Elixir/Erlang expert instead of having experts in Redis, cloud native, ... For better or worse, this is likely a large obstacle for many orgs.
The argument for using Elixir/Erlang is also more difficult when you have large companies like Github and Shopify demonstrating that Ruby can scale.
EDIT:
Let's not forget that it's mostly the DB that slows CRUD apps down. Not the language or the framework.
It'd be interesting if Mozilla offered a iCloud-like solution. Obviously, they wouldn't have the tight integration compared to Apple on mobile or desktop. They'd also need a really compelling use case that Apple can't offer.
What I miss most about LAN parties was the fun of getting all the computers setup. Setting up servers to host things for others to downloads, troubleshooting networking, random projects, ...