> you must go way out of your way to ignore how containerization makes everyone's job easier and even trivial to accomplish
You'd have to go out of your way to ignore how difficult they are to maintain and secure. Anyone with a few hours of experience trying to design an upgrade path for other people's container; security scanning of them; reviewing what's going on inside them; trying to run them with minimal privileges (internally and externally), and more, will know they're a nightmare from a security perspective. You need to do a lot of work on top of just running the containers to secure them [1][2][3][4] -- they are not fire and forget, as you're implying.
This one is my favourite: https://cheatsheetseries.owasp.org/cheatsheets/Kubernetes_Se... -- what an essay. Keep in mind someone has to do that _and_ secure the underlying hosts themselves for there is an operating system there too.
And then this bad boy: https://media.defense.gov/2022/Aug/29/2003066362/-1/-1/0/CTR... -- again, you have to do this kind of stuff _again_ for the OS underneath it all _and_ anything else you're running.
[1] https://medium.com/@ayoubseddiki132/why-running-docker-conta...
[2] https://wonderfall.dev/docker-hardening/
[3] https://www.isoah.com/5-shocking-docker-security-risks-devel...
[4] https://kubernetes.io/docs/tasks/administer-cluster/securing...
They have their place in development and automated pipelines, but when the option of running on "bare metal" is there you should take it (I actually heard someone call it that once: it's "bare metal" if it's not in a container these days...)
You should never confuse "trivial" with "good". ORMs are "trivial", but often a raw SQL statement (done correctly) is best. Docker is "good", but it's not a silver bullet that just solves everything. It comes with its own problems, as seen above, and they heavily outweigh the benefits.
> Explain in detail how being able to run services by running "docker run" is "overkill". Have you ever went through an intro to Docker tutorial?
Ah! I see now. I don't think you work in operations. I think you're a software engineer who doesn't have to do the Ops or SRE work at your company. I believe this to be true because you're hyper-focused on the running of the containers but not the management of them. The latter is way harder than managing services on "bare metal". Running services via "systemctl" commands, Ansible Playbooks, Terraform Provisioners, and so many other options, has resulted in some of the most stable, cheap to run, capable, scalable infrastructure setups I've ever seen across three countries, two continents, and 20 years of experience. They're so easy to use and manage, the companies I've helped have been able to hire people from University to manage them. When it comes to K8s, the opposite is completely true: the hires are highly experienced, hard to find, and very expensive.
It blows my mind how people run so much abstraction to put x86 code into RAM and place it on a CPU stack. It blows my mind how few people see how a load balancer and two EC2 Instances can absolutely support a billion dollar app without an issue.
> You're just publicly stating your ignorance. Do yourself a favor and check Ubuntu's microk8s. You're mindlessly parroting cliches from a decade ago.
Sure, OK. I find you hostile, so I'll let you sit there boiling your own blood.