This Week in DevOps | Hacker News Reader

This Week in DevOps

thisweekindevops.com

38 pointsciguy16 comments

I think this kind of sites are valuable as it is quite hard to keep track of the ever changing IT world.

In the beginning of this year we started https://nativecloud.dev/ which is a weekly curated list of noteworthy articles and tools of the cloud native world. We aim to publish every sunday.

I also recommend https://devopsish.com/ which is always on point.

DevOpsIsh is great, I am a subscriber there also. I started ThisWeekInDevops to serve a slightly different need for a similar audience. I wanted to keep up with just new announcements in the Cloud/DevOps space but without filtering all the content marketing that passes for announcements these days.

There was nothing out there quite like what I wanted so I started it myself. I focus on really only covering actual NEW stuff and filtering out the existing stuff.

I couldn't agree more!

This is why I started https://OpenSourceWeekly.org

The last issue is precisely about content curation :) https://opensourceweekly.org/issues/8

I like this. I once heard an anecdote from an AWS solution architect that they’re launching products so fast, they have real problems keeping up with it and keeping their staff trained.

The situation for us, however, is even worse: we’re expected to keep up to date with all these products, ideally across all cloud platforms. I find it simply impossible to keep up, and I appreciate this.

Thank you for the positive feedback, I really appreciate it! Creating this newsletter takes quite a bit of time expecially when there are a lot of announcements so it's great to hear that it's useful for someone besides myself.

A good example of this is AWS’ own Cloud Practitioner curriculum course topics versus what actually appears in the Cloud Practitioner exam. I scored a 100% on the knowledge check at the end of the course but a 70% on the proctored exam after missing quite a few questions about services that weren’t discussed in the study materials because the questions were covering comparatively new AWS services compared to what was in their own training material.

I've heard that it goes the other way, too: If you actually use AWS day-to-day, then you can have trouble on the exams because the exam is months out of date. So ex. if a service has some shortcoming and then they fix it, the exam might still expect you to say that the service can't do X.

>I find it simply impossible to keep up

How so? There are so many blogs, podcasts, news feeds, and youtube channels about clouds.

https://aws.amazon.com/blogs even have a RSS feed for lazy persons

I would say you have probably answered your own question: There are so many blogs, podcasts, news feeds, and youtube channels that it often feels hard to stay on top of exactly what is relevant.

Blogs aren't training material. Blogs are a story arc, and story arc TV shows tend to have problems courting new viewers in later seasons. People drop in, have absolutely no idea what's going on, and leave.

Anecdotally, I've met a couple at the party held by a common buddy who were sales managers at IBM and brought them the news that their company has a public cloud which went GA 6 months ago.

Happy accident, or on purpose?

If you can't even keep up with one vendor, then you'll have to choose one to pay attention to and give up on tracking the rest...

> The announcement does not specify whether this will also work with EKS (Kubernetes) Fargate but the mounts appear to be on the container level. Given this it seems likely that this will work for EKS also. If you have definitive information about this please let me know and I will send an update next week.

If I remember correctly, EFS is distributed NFS. If they support automatically provisioning AWS EBS blocks, mounting them into a fargate instance, and specifying IOPS/transfer speed/replication they'd surpass most of the offerings from other clouds platforms. Provisioning local/distributed storage (fs & block) is a massive pain in most clouds. AWS is one of the few clouds that I've used that let you provision IOPS and bandwidth. If you can provision block, file system, distributed, local, and specify performance requirements you'd be able to make some really efficient systems with no external DBs.

edited: typo NTFS->NFS

> EFS is distributed NTFS

I'm 99.999% sure this is a straightforward typo, but it's NFS (Network File System) not NTFS (the Windows NT File System).

Yep, too early and too little coffee. s/ntfs/nfs/g

You are correct on most points. I am actually looking at spinning up a basic EKS Fargate service to see if I can mount an EFS file system there also.

Reply on news.ycombinator.com