How to prepare for an interview with “AWS scenario questions”?

I have an interview for an SRE role tomorrow, I had the initial informal interview with them last week and it went quite well.

I got asked about prev experience and what I want to do next, as well as finding out about the company themselves.

Next up I have the second and final interview. Luckily I was told by the Head during the first interview that he’s not gonna make me do whiteboard tasks or make me code in front of them as he hates doing stuff like that, so why should he make me/us do it.

This was a relief, but the recruiter recently told me to expect some AWS scenario questions and some technical questions, but I don’t know what this means.

For context I’ve only worked one place so far in my 3 year career, I have applied for promotions and interviews tend to go as such:

- 60m technical exercise where I pair with a senior engineer and attempt a task like creating an sqs queue with boto3 and putting items through it, or making a lambda with terraform and trying to get it to respond to requests (we are told these technical tasks are not to see how “well” we can complete the task, but instead to see how we work and pair)
- 30m Q&A with questions asking for stuff such as benefit of DevOps, cloud computing, agile. Questions on working as a team, mentoring (basically behavioural based questions)

The above interview format is the only thing I’m used to as I’ve only ever worked in the same place, and now I’m looking for other jobs I’m not sure what exactly to expect.

I’m worried these aws scenario questions will be stuff like “here is X scenario, how would you build aws infrastructure to solve this” or “here is some aws infra, what is wrong with it and what would you change” which would instantly put me on the spot as usually with that sorta thing I’d be used to going off and googling to see what solutions and options there are.

Any advice? Thank you

https://redd.it/13oughj
@r_devops

Companies hire freshers as a DevOps Engineer role ?

I want to join as a DevOps Engineer role in IT field. Companies hire freshers or need some experience in any field then they hire ?

https://redd.it/13paj1n
@r_devops

How would you tackle dynamic code plugins

Hey there, first post here!
I’m currently tasked with a project to get an older application deployed via containers and Kubernetes. Full disclosure, I realize none of this is ideal.

So to put it simply, we have a PHP application that has a modular plugin directory where modules can be added and enabled for certain accounts instances of the application. The main application is now running in kubernetes however it’s modules are being tackled as a different beast. In total I’d say there are about 80 plugins total.

Here are some of key details and requirements

- dynamic updates of modules without needing to update the main app deployment/pods.
- ability to map the plugins to certain versions of the main app.

The plugins are developed separately in for repos with predictable name prefixes.


I’m currently debating the best approach for this with the following ideas.

1. Create a plugin manager app and deployment that grabs all plugins with matching prefix and downloads them to a pv backed by nfs and keeps them in-sync via release tags and the mounting the pv to the main app.
2. Build a container with all plugins and using it as a sidecar to populate the plugins directory.
3. Populating NFS via vía ci/cd by first deploying to object storage and running a sync of object to nfs. This would then be mounted to each pod as an nfs volume directly.
4. using an unit container to deploy the modules to the correct location.

If you were architecting this, how would you do so?

https://redd.it/13pblgd
@r_devops

Looking for hassle-free installation of Cilium on Kubernetes?

Check out our latest blog post for quick tips and configurations that will make your life easier. From IPAM configuration to LoadBalancer and Tunnel modes, we've got you covered!

https://medium.com/p/17a870fdc4f2

\#Cilium
\#Kubernetes
\#networking

https://redd.it/13pbxmz
@r_devops

How do you handle API documentation and change logs?

I’ve just gotten the go ahead to build out our companies first set of pipelines for our web applications. Unfortunately, I’m a bit unsure how API documentation and change logs should be handled/automated. One of the main goals with these new pipelines is to standardize and automate various parts of our development process so I’m essentially working from a clean slate.

Any advice or suggestions would be greatly appreciated.

General tips and things to look out for when building out CICD would also be much appreciated.

Thanks in advance!

https://redd.it/13pdlvn
@r_devops

Good Read: Guide to making on-call and managing incidents Zen!

Hahaha some one created this for our community. Kudos to this guy:P


"Inspired by Tim Peters’ The Zen of Python"


Order is better than disorder.
Positive is better than negative.
Straight is better than vague
Communicating now is better than communicating later.
An incident should never go unacknowledged.
Unless explicitly supressed.
An incident has one and only one commander.
Context counts.
So does documentation.


Read full version here:

https://www.zenduty.com/blog/the-zen-of-on-call/

https://redd.it/13pg6q7
@r_devops

Multi Stage Dockerfile Examples In Node.Js?

I searched but could not find any tutorials for it. Can you please share them?

https://redd.it/13pe38m
@r_devops

Struggling to understand difference between build merge pr deploy release

I see these terms thrown around and try my best to understand them. What are these steps and in what sequence they occur ? Where can I read more about them ? Thank you.

https://redd.it/13pdi6y
@r_devops

Ok, this is dumb but... Anyone tried to back up and restore a VM with active containers on it?

Do the containers... handle it well? Assuming it's a snapshot backup of the whole OS

https://redd.it/13osegr
@r_devops

How to use multus CNI and load-balancing

A good blog about this topic : https://cloudybytes.medium.com/k8s-bringing-load-balancing-to-multus-workloads-with-loxilb-a0746f270abe

https://redd.it/13on9rn
@r_devops

Overview of DevSecOps Tools to Secure your Applications in 2023

By integrating security into the entire software development process, DevSecOps tools identify its applications’ vulnerabilities and improve overall application security. Find the coplete list of DevSecOps tools here.

https://redd.it/13plqgd
@r_devops

The Tezos Art x Web3 Hackathon Will Be Held At The Paris Ubisoft Headquarters

During the 3-day event, participants will create evolving / interactive Web3 on-chain artworks.

In the hackathon 6 teams will compete, these 6 teams will be formed around 6 top international digital artists.

You can read the article in full below : ⬇️

https://xtz.news/adoption/the-tezos-art-x-web3-hackathon-will-be-held-at-the-paris-ubisoft-headquarters/

https://redd.it/13pladb
@r_devops

Any exciting projects/tools

fed up with all AI, terraform noise. what are you excited about ?

https://redd.it/13pnsa7
@r_devops

When do you promote from Dev->QA->Stage->Prod?

As per title. I'm curious what's the criteria at your company, when do you move from an environment to the other? Also, how do you automate this, or does it require manual intervention?

Curious to see how other team and companies' practices.

https://redd.it/13ppz4i
@r_devops

How to add non modular jar in module project ?

stackoverflow-source

I tried import junit.awtui in my module project main class

Error report : The package junit.awtui is not accessible

Problem is junit jar hasn't module-info.class, so its packages are not accessible
How can I resolve the problem ?

https://redd.it/13prgrs
@r_devops

DevPod Demo - Walkthrough of the UI and CLI tips (Open Source alternative to Codespaces)

Here's a demo with one of the DevPod creators. We walk through how to create workspaces, how providers work, how the CLI can be used instead of the UI, and other topics we haven't covered previously in videos.

If you have questions, or want to see content around some of the topics we briefly touched on, let us know!

https://youtube.com/live/Bu-aGpzwMUw

https://github.com/loft-sh/devpod

https://redd.it/13psb7z
@r_devops

Jenkins credentials

So, I was very excited about creating my first pipeline from scratch! Using Terraform to write my code and a GitHub repo, then now it comes to automating with Jenkins and I keep getting this error message



Error: configuring Terraform AWS Provider: failed to get shared config profile, default with provider["registry.terraform.io/hashicorp/aws"\], on providers.tf line 9, in provider "aws": 9: provider "aws" { Build step 'Execute shell' marked build as failure Finished: FAILURE



The code runs with no errors from the Cloud9 instance but for some reason, Jenkins cannot access the shared credentials file. Help, please

https://redd.it/13pux6j
@r_devops

Learn how elastic.co "does" internal eng productivity this Friday

This Friday, Stacey King Poling, Director of Engineering at Elastic.co will discuss eng productivity as part of Aviator's "off the record" series.

It's free and not recorded. Hope to see you there.

https://twitter.com/Aviatorco/status/1661070867453886465?s=20

Sign up: https://dx.community/

https://redd.it/13pw9hy
@r_devops

dagger tool

Hello, folks!

I have been studying a bit the dagger for CICD tool at https://dagger.io/
For me it's pretty cool how pipelines could be created using Python/others programming languages instead of using YAML / pwsh/ bash scripts/ GitHub actions.
Other positive point is to have a way to test the pipelines without spending time on cloud CICD solutions.
What do you think about? Are you currently using for your projects/in your company?

Appreciate your comments

https://redd.it/13pvqlc
@r_devops

Hypocritical managers and devops

Has anyone else noticed the number of managers that are absolute hypocrites? They want us to do strict 9 to 5, in the office a certain amount. And then they are barely visible from the org standpoint, work from home without any schedule for us to know when they'll be available. I'm at the point where I've been with this org for a year now almost and I think I'm ready to move on already. Like if I had kids it'd be a lot easier to get him to be flexible about the hours. But since I don't have kids it's like he expects my entire life to revolve around this team.

Like, I was 7 years in the industry before this job. Have appointments, health issues, have PT, etc. Those are ALL during the day. It's weird, as long as he doesn't see or is doing his own thing he doesn't care. But he just gets these random moments of "i must exercise my power and feel powerful" and we're just supposed to bow down. It's so cringe and I'm ready to leave this employer.

/vent /rant

https://redd.it/13pwwnb
@r_devops

Help me understand trunk based development (for academia and research usefulness)

I'm an academic researcher. I have a deep fear of one day being one of those old professors who still thinks LabView is the greatest thing since sliced bread and that there isn't even a reason to give something like Python the time of day.

Git flow is an excellent branching model. It still has its place for things like open-source projects where you need to control things tightly. It is also relatively simple, and can easily be explained and followed by other academics and scientists who, while they write good code for their domain, are not themselves software engineers.

I'm a proponent of semantic versioning and conventional commits. I also believe deeply in achieving 100% code coverage and that this plays a significant role in continuous integration before allowing merges back into main.

I keep hearing about trunk-based development. I've read many articles. I've read most of trunkbaseddevelopment.com. However, it all seems like a lot of talk. It looks even more so to be just the absolute Wild West and an excuse to develop however you want. Of the few more fleshed-out examples I have found, it quite honestly seems to largely resemble git flow in many ways, minus the develop and release branch, just going straight from feature to main, assuming it has passed all the CI checks.

Emphasis keeps getting placed, it seems, on these short-lived feature branches so that you can avoid merge hell. Merge hell is not a consequence of long-lived feature branches. It is a consequence of inadequate project management, in my experience. It stems from needing a proper issue tracker and project manager; from two people or groups working on the same area of code simultaneously.

I understand trunk-based development is not meant to replace things like git flow for open source, but I am interested in if it fits internal projects. But I am seriously failing to understand the appeal or, quite frankly, the point.

I especially don't understand this rapid or short-lived feature branch talk. Sometimes features take a long time to develop. Writing good unit tests and integration tests often takes longer than writing the code itself. There's no getting around that and no changing the fact that you shouldn't be prematurely merging code from a significant addition that will take time to engineer out properly.

I don't understand what trunk-based development actually solves, and that isn't just a consequence of poor project management.

What am I missing or failing to understand here? Are there any good articles or resources that talk about and address these kinds of issues in a salient way that are not just pure hype clickbait blog posts?

Thanks in advance.

https://redd.it/13pyae7
@r_devops