Skip to main content


I'm watching a video training on Kubernetes. It's explaining the advantages of abstracting container management (which abstracts application management (which abstracts data and algorithms (which abstracts hardware...)))

... And then says it makes deployment more efficient.

What does efficiency mean to these people?

reshared this

Unknown parent

Jess👾
God help you if your pacemaker is running Kubernetes!
Unknown parent

spiralbatross
that runs in my heart… uh, wait…
in reply to FoolishOwl

usually it means "someone else has to deal with this problem that isn't the engineer building your application layer".
in reply to FoolishOwl

Ye gods, the next section of this course discusses an abstraction layer for management of Kubernetes pods.

I can't feel my toes.

in reply to FoolishOwl

honestly, I’ve used docker containers a ton, and I’ve occasionally tried to look at using kubernetes… and everytime I immediately bounce off, unable to even get to wherever they describe what the fuck it’s doing, much less what situation it’s supposed to simplify.

I still have no idea. Scaling, I assume.

in reply to Asta [AMP]

@aud I think, yes, scaling.

Most of what I know about Docker is that it's ubiquitous for running services on an SBC. It made it relatively easy to set things up with a reverse proxy, but I've wondered if it would be worth figuring out how to run the services without the containers for increased efficiency.

in reply to FoolishOwl

OMFG. The next section after that discusses GKE, yet another layer of abstraction to manage the Kubernetes management tools.

This is described as "efficient" three times in three minutes.

in reply to FoolishOwl

Nominally this is a series of courses on using Python for automation. The original problem, supposedly, was making sure you had the libraries available for your Python script.
in reply to Alex P. 👹

@saddestrobots It's the fifth course of a series of six on Coursera. This course started by talking about how a developer might bundle code in a Docker image, and then said a Python developer, apparently as an afterthought. It was a very confusing presentation and didn't even explain how to use a Docker image.

The other courses in the series were at least quick introductions to practical skills. I think this one is basically advertising Google's cloud services.

in reply to FoolishOwl

As best I can make out, they keep using "efficiency" to mean specifically minimizing time to deploy code.

It's like saying a helicopter is more efficient than a bicycle, because it's faster.

It's ignoring all the other costs that increase as you focus on optimizing for speed of deployment.

in reply to FoolishOwl

hmm I would not describe GKE as an abstraction, I would describe it as k8s clusters as a hosted and managed service. You can use all the standard k8s interfaces to actually manage the containers once the cluster is set up
⇧