Azure and Aws interview questions

Hi all my friends at ireland trying for cloud and devops freshers role if you have any questions dump share here
Thanks in advance.

https://redd.it/1ot08j2
@r_devops

How to do ci/cd on an api? stuck with intuition of multi local/staging/prod codebases

Hi guys, I built a nice CI/CD pipeline for an app -- took me a while to learn, but it now makes intuitive sense with local/staging/prod. You push small commits and it auto-deploys. That makes sense when you just have that one pipeline.

But now, how do you apply that to an API? By design, APIs are more stable -- you aren’t really supposed to change an API iteratively, because things can later depend on the API and it can break code elsewhere.
This applies to both internal microservice APIs (like a repository layer you call internally, such as an App Runner FastAPI that connects to your database --/user/updatename), and to external APIs used by customers.

The only solution I can think of is versioning routes like /v1/ and /v2/.
But then… isn’t that kind of going against CI/CD? It’s also confusing how you can have different local/staging/prod environments across multiple areas that depend on each other -- like, how do you ensure the staging API is configured to run with your webapp’s staging environment? It feels like different dimensions of your codebase.

I still can’t wrap my head around that intuition. If you had two completely independent pipelines, it would work. But it boggles my brain when two different pipelines depend on each other.


I had a similar problem with databases (but I solved that with Alembic and running migrations via code). Is there a similar approach for API development?

https://redd.it/1osyf94
@r_devops

Building a CI/CD Pipeline Runner from Scratch in Python

I’ve been working with Jenkins, GitLab, and GitHub Actions for a while, and I always wondered how they actually work behind the scenes.

After digging deeper, I decided to build a CI/CD pipeline runner from scratch to truly understand how everything operates under the hood.

As DevOps engineers, we often get caught up in using tools but rarely take the time to fully understand how they work behind the scenes.

Here’s the full post where I break it down: Building a CI/CD Pipeline Runner from Scratch in Python

https://redd.it/1ot3a9c
@r_devops

Infrastructure considerations for LLMs - and a career question for someone looking to come back after a break?

This sub struck me as more appropriate for this as opposed to itcareerquestions - but if I'm off topic I'm happy to be redirected elsewhere.

I've 20+ years working in this kinda realm, via the fairly typical helpdesk - sysadmin - DevOps engineer (industry buzzword ugh) route.

I am the first to admit, I very much come from the Ops side of things, infra and SRE is more my realm of expertise... I could write you an application, and it'd probably even work, but a decent experienced software developer would look at my repo and go "Why the feck have you done that like that?!".

I'm aware of my stengths, and my limitations.

So... Mid 2023 I was made redundant from a ",Senior Managing DevOps consultant" role with a big name company known for getting a computer to beat a chess grand-master, inspiring the HAL-9000 to kill some astronauts (in a movie), kmown for being big and blue...

70,000 engineers got cut. Is what it is. Lots of optimism about AI doing our jobs, some mixed results.

I took a bit of a break from the tech world, professionally anyway... I actually took on managing a pub for a year or so. Very sociable, on my feet moving around... I lost a lot of weight, but not good for my liver, I had a lot of fun... Mayhe too much fun.

Now - I'm looking at the current market, and reluctantly concluding, the thing to do here is become proficient at building and maintaining infrastructure for LLMs...

But my google (well duckduckgo) searches on this topic have me looking all over the place at tools and projects I've never heard of before.

So - hive mind. Can anyone recommend some trustworthy sources of info for me to look into here?

I am fairly cloud savvy (relatively) but I have never needed to spin up an EC2 instance with a dedicated GPU.

I am broke, like seriously broke...my laptop is a decade old and sporting an I5-2540M. I am kinda interested in running something locally for the exercise of setting it up, fully aware that it will perform terrible...

I don't really want to go the route of using a cloud based off the shelf API driven LLM thing, I want to figure out the underlying layer.

Or, acknowledging I am really out of my element, is everything I'm saying here just complete nonsense?

https://redd.it/1ot26kr
@r_devops

How can i host my AI model on AWS cheap ?

Sorry if this comes as dumb. Im still learning, and i cant seem to find an efficient and CHEAP way to get my AI model up n running on a server.

I am not training the model, just running it so it can receive requests

I understand that there is AWS bedrock, sagemaker, avast AI, runpod. Is there any cheaper where i can run only when there is a request ? Or i have no choice but to get an ec2 to constantly run and pay the burn cost

How do people give away freemium for AI when its that pricey ?

https://redd.it/1ot5ukb
@r_devops

What’s your go-to API testing tool in 2025 for CI/CD pipelines?


Hey everyone,

Our team’s been revisiting our API testing and documentation setup as we scale a few services, and we’re realizing how fragmented our toolchain has become. Postman’s been reliable, but the pricing and team management limits are starting to hurt.

We’re evaluating newer or lighter tools that integrate well into CI/CD workflows ideally something that handles API testing, mocking, and maybe documentation generation in one place.

Here are some we’ve looked at so far:

- Katalon – lots of automation features but feels heavy
- Hoppscotch – nice UI, but limited for team workflows
- Apidog – looks interesting since it combines testing + documentation and supports API collaboration
- Insomnia – still solid, though team features are a bit clunky
- Bruno – nice offline Postman-style tool

Would love to hear from others what’s been working well for your devops/testing teams lately?
Anything that actually fits into CI/CD pipelines cleanly without 20 different integrations?

https://redd.it/1ot8f0x
@r_devops

I am building a lightweight engine for developing custom distributed CI/CD platforms. It makes building and managing custom CI/CD platforms easier by handling the orchestration so you can focus on how your workflow works

Leave a github star, if you find the project interesting. https://github.com/open-ug/conveyor

https://redd.it/1ot967b
@r_devops

How would you set up a Terraform pipeline in GitHub Actions?

I’m setting up Terraform deployments using GitHub Actions and I want to keep the workflow as clean and maintainable as possible.

Right now, I have one .tfvars file per environment (tfvars are separated by folders.). I also have a form that people fill out, and some of the information from that form (like network details) needs to be imported into the appropriate .tfvars file before deployment.

Is there a clean way to handle this dynamic update process within a GitHub Actions workflow? Ideally, I’d like to automatically inject the form data into the correct .tfvars file and then run terraform plan/apply for that environment.

Any suggestions or examples would be awesome! I’m especially interested in the high-level architecture

https://redd.it/1otaesy
@r_devops

has ai actually improved how you code?

i’ve been using chatgpt for a while and added cosine recently for my personal python projects. it definitely makes me faster, with cleaner code, quicker debugging, and better structure, but sometimes i feel like i’m getting too reliant on it.

i’ve noticed that ai tools can speed up routine work, but when i hit a problem that needs deeper thinking or system-level decisions, i catch myself opening chatgpt instead of figuring it out myself.
it’s great for productivity, but i’m not sure if it’s actually making me better at problem-solving in the long run.

curious what others in the industry think. has ai genuinely improved your technical skills, or are we just becoming better at prompting and outsourcing the hard parts?

https://redd.it/1otc7gz
@r_devops

KubeGUI - Release v1.9.1 dark mode, resource viewer columns sorting and large lists support

🎉[Release\] KubeGUI v1.9.1 - is a free lightweight desktop app for visualizing and managing Kubernetes clusters without server-side or other dependencies. You can use it for any personal or commercial needs.

The items we discussed before are now being introduced:

+ Dark mode.
+ Resource viewer columns sorting.
+ All contexts now parsed from provided kubeconfigs.
+ On startup if local KUBECONFIG env var defined - contexts will be inserted automagically.
+ Resource viewer can now support large amount of data (tested on ~7k pods clusters).
+ Bunch of small ui/ux/performace bug fixes.

Kubegui runs locally on Windows & macOS (maybe Linux) - just point it at your kubeconfig and go.

\- Site (download links on top): https://kubegui.io

\- GitHub: https://github.com/gerbil/kubegui (your suggestions are always welcome!)

\- To support project: https://ko-fi.com/kubegui

Would love to hear your thoughts or suggestions — what’s missing, what could make it more useful for your day-to-day ops?

Check this out and share your feedback. ps. no emojis this time! Pure humanized creativity xD




https://redd.it/1otbphi
@r_devops

How to stay updated and keep upskilling.


I have been in devops role from last 1 year. I was dealing with docker, linux machines on aws and linode. It was a small scale startup they had around >20k daily active user. I have resigned in sept as i needed a long break (4 months) due to some personal work. Currently i am a bit worried what if i forget how to do this that stuff in devops. I just wants to know how can i keep my self aligned with the market so if i start job hunting after my break i don't feel under skilled. How to practice devops on scale to keep the confidence.

Thanks

https://redd.it/1ote087
@r_devops

VOA v2.0.0 — Secrets Manager

I’ve just released VOA v2.0.0, a small open-source Secrets Manager API designed to help developers and DevOps teams securely manage and monitor sensitive data (like API keys, env vars, and credentials) across environments (dev/test/prod).

Tech stack:

FastAPI (backend)
AES encryption (secure storage)
Prometheus + Grafana (monitoring and metrics)
Dockerized setup

It’s not a big enterprise product — just a simple, educational project aimed at learning and practicing security, automation, and observability in real DevOps workflows.

🔗 GitHub repo: https://github.com/senani-derradji/VOA

you find it interesting, give it a star or share your thoughts — I’d love some feedback on what to improve or add next!


If

https://redd.it/1otf1zr
@r_devops

In 2022, I wrote that DevOps had become waste, in 2025 AI is the new waste!

In 2022, I said DevOps had become waste.

The response?
"DevOps can't be waste we need automation!"
They missed the point.

DevOps principles were right.
But when every team rebuilds the same CI/CD pipelines, writes the same Terraform modules, and solves the same problems in isolation
that’s not DevOps.
That’s local optimization at scale.

Now it’s 2025. AI is the new waste.

Team A spends two sprints wiring up Claude to “understand” their codebase.
They chunk it, inject docs, tweak prompts.
Team B? Doing the same thing.

Different team. Same half-baked playbook.
No shared learning. No standardization. No outcomes tracked.

And most orgs?
Still stuck trying to pick Copilot vs. CodeWhisperer vs. Windsurf
with zero plan to measure impact or build repeatable systems.

This is Jenkins sprawl all over again but for cognition.

I call the fix: OutcomeOps
https://www.outcomeops.ai/blogs/outcomeops-ai-is-the-new-waste

https://redd.it/1oti8e8
@r_devops

CKA Preparation

Im preparing for the CKA Cert. I already did these courses: LFS158 & LFS258, and I’m administering the k8s cluster of my company for a little more then a year now on pretty much a daily basis. I did the killerkoda tests & also did both of the killer.sh mock exams. In the first mock exam, I only scored about 50% and in the second one even worse. I used the 120min timer to make the test as realistic as possible. After this I redid all of the answers that I failed on & got 100% correct. I didn’t really have issues with specific topics, my only problem was the time constraint.
So my question: Am I prepared enough, even though I technically failed the mock exams? I read that killer.sh exams are much harder then the real exam. If that’s not true, I don’t really know how to better prepare for the exam, because I prepared using all of the resources that I’m aware of.

Thanks :)

https://redd.it/1otedcw
@r_devops

How do you check or enforce code documentation in your pipelines (C/C++ & Python)?

Hey,

Currently working on improving how we enforce code documentation coverage across a few repositories, and I’d love to hear how others handle this.

We have three main repos:

one in C++
one in C and C++
one in Python

For C and C++, we’re using Doxygen with Javadoc-style comments.
For Python, we use Google-style docstrings.

Right now, for the C and C++ part, we have a CI pipeline that runs Doxygen for every merge request and compares the documentation coverage against the main branch. If coverage decreases, the user gets notified, and the MR is blocked.

That works okay, but I’m wondering:

Are there better or existing tools or CI integrations that already handle documentation checks like this? Only Open source and applying locally would be fine.
What would be a good equivalent setup for Python? (e.g., something to validate or measure docstring coverage)
Has anyone implemented pre-commit or pre-push git hooks that check for missing documentation or docstring issues before the MR even gets created?

Thanks in advance!

https://redd.it/1otjf9r
@r_devops

Open-Source ACME server - 100% CertBot compatible - One binary

Hi everyone!

We have developed an Acme server for our use case. It is written in Rust, which means you only need to work with a single binary. In file mode, our test is 100% compatible with the existing Certbot solution.

For more details, visit: https://github.com/arxignis/ssl-storage

**Summary:**

✅ Written in Rust

✅ Fully compatible with Certbot

✅ Utilizes a Redis backend for storage

✅ Supports distributed mode (with Redis)

✅ 100% compatible with CertBot

✅ Redis backend as storage mode

✅ Distributed mode (with Redis)



https://redd.it/1othpt8
@r_devops

Browsing helm chart from terminal - LazyHelm


Hi community!

Sometimes, when I deploy or test some application, I prefer looking into helm charts using directly the terminal and I found using helm commands alone can get a bit tedious, so I tried to created something to make it easier.

So I tried to create (with ai helps) something that makes the process easier, LazyHelm.

It’s a small personal project I built to make my own workflow smoother, but I hope it might help someone else too.

What it does:

Organized menu system to browse local repositories or search Artifact Hub
Browse your configured Helm repos and discover all available charts
Find charts across Artifact Hub directly from the terminal
Add, remove, and update repository indexes with simple keystrokes
Inspect chart values with syntax highlighting and diff between versions
Modify values in your preferred editor ($EDITOR) with YAML validation
Fuzzy search through repositories, charts, and values
Copy YAML paths to clipboard or export values to files

All in your terminal. No need to remember helm commands or manually fetch values.

Installation via Homebrew:

You can install LazyHelm using Homebrew:

- brew install alessandropitocchi/lazyhelm/lazyhelm

GitHub: https://github.com/alessandropitocchi/lazyhelm

Any feedback, suggestions, or feature requests are very welcome!

Thanks for reading!




https://redd.it/1otopeq
@r_devops

QA team was cut in half, facing the same release pressure. thoughts?

we lost half of our QA team in the last round of budget cuts, but somehow leadership is still expecting us to keep shipping every 2 weeks. I mean manual regression alone takes most of the sprint, not to mention the pain of cross device tests as we're testing across web + android.

the team is already burned out and lacks resources now, higher ups say we can fix this with automation but setting up new frameworks feels like starting a new project and we can't afford to waste any more time experimenting nor do we have the engineering bandwidth now...

has anyone successfully automated testing across devices without hiring more engineers? AI tools? Low-code? we need something good and we need it SOON...

https://redd.it/1otts8b
@r_devops

Moving to a mid level position

Hey all,

So, I've been within the devops/platform engineering space for just under 2 years now. I come from a non tech background but I'm firmly in the tech space now.

But I wanted to understand how can I make that move from junior to mid level engineer? I have a good solid grasp of Terraform, GitLab CI. Some Docker and K8s skills (fairly new for a project on EKS). My main cloud is AWS for the past 3 years. I'm currently also getting involved with some other clouds like oci.

But I feel like I don't have a strong understanding of some basic stuff that an IT or tech guy should have. Networking skills are probably lacking tbh. I'd love to increase my security skills also.

I would love to have someone as a mentor to help guide and advise me through this process.

https://redd.it/1otu4z2
@r_devops

Policy as Code

I recently moved our company’s azure policy away from being manual process through the azure web portal to a pipeline using terraform. It’s working but it’s not great, I’m wondering how others manage their Azure Policy, or AWS scps

https://redd.it/1otyhkh
@r_devops

HTTP Parameter Pollution: Making Servers Disagree on What You Sent 🔀

https://instatunnel.my/blog/http-parameter-pollution-making-servers-disagree-on-what-you-sent

https://redd.it/1otzhyy
@r_devops