Best way to learn scripting languages

Hello guys, looking for some advice. I'm into DevOps/SRE.

My current project or even past one's never had much to do with coding.
And now I want to switch my job, and I want to learn some scripting with Python and bash.

What is the best place to start learning these skills? My current project ends in 3 months, and I need to prepare for scripting as it's very common in all the JD's.

Additionally I would like to know what automation/scripting/coding you guys are doing on a daily basis? Would be helpful if you mention some tasks or some scenarios to prepare for the interviews.



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

Looking for a serious learning partner AWS, DevSecOps, Azure, Kubernetes

Hey everyone, I’m looking for a dedicated learning partner to study Skills above mentioned together. My goal is to stay consistent, exchange knowledge, and keep each other accountable. Ideally, we can set up regular check-ins, discuss concepts, and motivate each other.

A bit about me: I have almost 4 years of experience, but I feel my knowledge is scattered and not at the level I want it to be. I’m looking to improve my skill set in a structured way and would love to learn alongside someone with the same mindset.

If you’re interested, feel free to comment or DM me. Let’s grow together!

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

Stay at current job or go for in-office a position?

Hey folks, posting this on a burner :) I'm conflicted in my current situation and would love some opinions on what company you would choose?

Company A (Current) - Contract - Remote - Large Media company - Sr.SRE

$90/hr \~ $175k yearly
Ok benefits for Medical, Dental, Etc
No PTO or paid holidays
No 401k or retirement options
Very chill workload
Contract until 2026 (Have been extended multiple time with likelihood of converting to FTE)

Company B - FTE - In-Office 5 days - Start up building a space station - Sr DevOps Engineer

185k + $5k bonus + equity
Fully covered benefits
Paid holiday and flexible PTO
1% 401k Match
Commute is 30 minutes 1-way
On-site perks such as free lunch, etc

Thanks in advance!



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

RPA vs AI Agents vs Agentic Process Automation – What’s the Future?

Hey everyone, I’ve been seeing a ton of posts lately about how RPA is either on its way out or evolving into something new with AI agents. Some people swear LLM-based agents are the future, while others think RPA isn’t going anywhere—it’ll just merge with AI to create these hybrid systems people are now calling Agentic Process Automation (APA).

From what I’ve read, APA is basically RPA that still handles repetitive rule-based tasks but also has some level of reasoning thanks to LLMs. Makes sense, but also kinda feels like buzzword soup.

Honestly, I’m lost. Is APA actually the future? Or are we just slapping new labels on the same concepts?

I’ve been working more with AI agents (LangGraph, AutoGen, CrewAI, etc.), but I keep hearing that these setups are expensive and that most companies will go for hybrid solutions instead. If that’s the case, should I bother learning RPA, or just keep focusing on AI agents?

Would love to hear from people who’ve been in the automation space for a while. Are we heading toward full AI, or will hybrid models dominate?

Appreciate any insights!

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

What's your take about on-call?

Been there, done that...hated it.

My first job was a kind of Helpdesk/SysAdmin role where I did it in a 24/7 base and had to wake up in 4 A.M from a cell phone ringing because a ship crew member from Philippines didn't had internet access (F.M.L).

This, among with me having different ambitions and some weired things that were happening at that company, brought me on switching to DevOps with which I'm pretty happy and I can clearly say that it was the right choice.

Although I see that nowadays the on-call thing is becoming a kind of a standard for DevOps with more companies seeking out for engineers that are willing to do it.

What's your take on that? Is it really a thing? Can you see it growing?

https://redd.it/1ituwh1
@r_devops

Need some help into deploying code created with codecraft.name - code to app - ai agent

Hello all d3vs, i am not myself a dev but i twinkle with my proxmox and i have a homelab ai workstation with 4 x 3090 , 64 core amd epyc and 256 gb of ddr4 ecc.
I come across a project called www.codecraft.name $CRAFT , and i start to get attached to it as i like the way is coding but i want to help the project and i was looking in some ways, hot to get the generated code from the website and create an ai agent with n8n and be able to deploy the app from code ?
I mean i know not all f4ameworks are easy to inplement but what is the easiest method/framework that an user can deploy his code and see his app/game working ?

It is possible to do it with n8n into something like automatic deployment ?

Much appreciated!

https://redd.it/1ity87b
@r_devops

I messed up - came here for lashings

We're still building out our environments and there were some things that were lower priority on our tiny team (entire group of 10 people). One of those things was putting in a codeowners file in most repos.

We have a reusable workflows repo where we put everything that's not a one off and other repos call those workflows. Anything that touches our actual infra or service outside of GitHub has federated credentials that are tied to the common workflow repo. Basically anything important has to go through the reusable workflows repo.

Yesterday I get pinged about some workflows failing. Which was interesting because nothing had been touched from our end.

I went and looked... One of the management team had told an intern to start building out their own workflows... Someone that has no idea what they're touching. And things were failing because they couldn't authenticate and other stuff I do have protected.

So today I'll be adding codeowners protection on my .github directories.

Please chastise me here for not doing this sooner and creating more work for myself.

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

Need advice for my CKA retake

Yesterday, I took the new CKA exam. I knew there would be some changes, but I brushed them off, thinking they would be minor (my dumb ass taking a cert for the first time). I got 67/125 on my first try at Killer.sh, every single question there took me forever. But after a few more attempts, I managed to clear it, so I felt confident and booked the exam.

However, the real questions completely threw me off: API Gateway, which was never mentioned in the Udemy course, installing packages using dpkg, and a few other surprises. Aside from those, everything else was quite basic and similar to the simulators, just worded differently.

By the end of the exam, I had 3 out of 16 questions left unanswered (one on API Gateway, one on Helm and dpkg, and one on setting up a cluster using kubeadm). A couple of answers I was unsure about, and another one or two were only partially completed (missing just one requirement).

I already have a preparation plan, like practicing setting up a cluster smoothly and preparing for dpkg, Helm, API Gateway, and network plugins. But in the worst-case scenario, it would still be a 50/50 chance. So I really want to prepare properly for the retake because waiting for results in that situation was awful.

Any advice on what I should focus on for my next attempt? Thanks!

https://redd.it/1iu0buv
@r_devops

Which processor should I choose for my VPS?

Hey there

Im a frontend dev, I built my app and hosted it on vps on nginx + docker. My current vps is a bit expensive and I want to change my vps provider.

I was thinking about swapping to hetnzer but im not sure which CPU should I choose for my vps.

there are 3 available options there:
1. Intel® Xeon® Gold.
2. AMD EPYC™ 7002 series 
3. Arm64 Ampere® Altra®

the arm64 is the best price to performance option.

What option should I choose for tools that I use?

My front app is react,
I have 2 rest api's - one is nodejs, second one is laravel
i have mysql db on my server
also i use docker and nginx

Could you guys please help me to choose the right option?

https://redd.it/1iu213j
@r_devops

Very clean linux commands cheatsheet

Link: linux.devopscommunity.in

https://redd.it/1iu3y0e
@r_devops

Technical Interview Round requires me to setup an AWS account with activated payment. Is this normal?

I got a schedule for a technical round where the requirement is I have my own AWS account with me paying for it, as well as terraform, kubectl, python and aws CLI all installed.

The most absurd this is the position delves into almost everything:

- Hardware Design and Simulation
- IoT Fleet Management
- Project Management
- Edge Computing
- Data Analysis

Ironically, the interview is scheduled with Cloud Setup, which is a bit baffling, there is never any mention of DevOps requirements in the position and it is just a Software Engineer Position with lead in brackets.

Is this common practice? Who is supposed to bear the cost during the interview and if something goes awry?


https://redd.it/1iu3vos
@r_devops

What would be your setup in order to manage Github repos at scale?

Hello!

In my company we are managing quite a big number of products, and each of them have their own repo.

The CICD Pipeline looks however the same modulo few changes.

How would you manage that from a CICD point of view? One way would be to build actions and reuse them in each repo, but at scale it might become a mess to redefine everything.

Do you have any ideas?



https://redd.it/1iu5vtg
@r_devops

Infrastructure as Code Survey

Take this quick (10 question) survey for a chance to win $20!

https://forms.gle/dffsFEkdFc1S2gYB6

We are performing a survey of the Infrastructure as Code landscape to learn more about how engineers are interacting with Infrastructure as Code tools, the technologies they prefer, and the services they rely on.

As a thank you for participating, we'll be randomly selecting 10 respondents to send $20 gift cards to.

Share with your friends!

https://redd.it/1iu8gb9
@r_devops

Devops Sales Executives?

Anyone on here in the sales side of things? Would love to connect with others who are selling Devops product and service.

https://redd.it/1iu7a0u
@r_devops

Terraform for intern

Hello i na intern in devops. I learn at job so i have a question. I ma in really big project with a lot of files in repo of terraform and i feel overwhelmed with this complexity. Do you have some tips how to understand this modules and why there is this type of input? Have a nice day

https://redd.it/1iua085
@r_devops

How do you organize tf projects?

I have a project that started with a setup project, this is where I’d define dependencies for everything like ecr repos, Iam etc.

Deploy having the infra for my initial project. Now I have like 3 projects that are all intermingled.

I’m thinking that within deploy I’d have shared resources, everything else would be broken down into project modules with their respective iac for ecs, rds, etc etc.

Any good references/ideas for mono repo structure?

https://redd.it/1iuc3ob
@r_devops

Recommend an automated build and deployment system for a small company

I've recently accepted a developer role at a (very) small company that sells a niche software product, in both SaaS and run-on-your-desktop variants. The company has been around for \~20 years, and all of their practices are from that era - EVERYTHING is completely manual, and done directly from developer machines, up to and including production deployments. There's little to no visibility of which software versions are running in which environment, no centralised repository for configuration, and so on.

There are only 3 IT people in the org - me, the dev who originally wrote the software, and an "IT Ops" guy who manages servers, databases, networks, and so on.

I've managed to sell the concept of automated builds and releases to management, and the next step is to write up a proposal including costs and benefits.

Where I'm now stuck is which automated build/deployment product to put into the proposal. The basic requirements are:

* Automated builds - codebase is 90% C#/.NET, with some exceptions - some C++ code for performance intensive stuff, and about half of the web code (Typescript / JQuery / React) is currently built using yarn.
* Support for \~30 applications, a few of which are software releases to customers but most of which are backend API's, web applications, or batch processing apps running on our (bare metal VM, no Kubernetes/Docker) infra.
* Support for a Windows-only environment, with apps running as a mix of console applications in the foreground, windows services, or web applications hosted in IIS.
* Ideally a simple UI showing a matrix of environments, applications, and software versions - something suitable for e.g. a product owner.
* Selectable versions and deployment targets with manual release triggers. We're a long (loooooong) way from true CI or CD. One-click stop/start/upgrade for our IT Ops guy, with dropdowns for app versions ideally driven from git tags in the associated repo (or similar).
* Email notifications of software releases to the broader team ("Application A version 3.x.y has just been released to VM1 in Production, release notes here <insert text from release notes file>").
* Constrained targets for each project - Application A should only be able to run on VM3 or VM5 in the prod environment, etc.
* Scriptable deployments, or even something e.g. YAML-based as long as custom plugins are possible in Python, C#, Powershell, etc.
* Affordable - our operating budget is low, as you might imagine.
* Simple and maintainable - we don't have a dedicated DevOps person, and our IT Ops guy isn't going to spend weeks or months traversing a steep learning curve.
* Eventually, support for automated tests and code quality checks - none of these exist right now and the codebase is a spaghetti mess, but that's something that will now be improving over time.

I'd previously stumbled across Octopus Deploy, which seemed to tick all of our boxes - but the recent price increases have now put it well out of our budget.

Any helpful recommendations gratefully received. And no, "find another job" isn't a helpful response in this instance :P. These folks are a joy to work for in many ways, just not this particular one - and at least they're open to improvement.

https://redd.it/1iubcm8
@r_devops

Cloud Provider that offers prepaid compute?

I want to host a pretty simple backend, in addition to a small sql database somewhere on the cloud. However I am worried to host this all on AWS or Google Cloud, as they ostensibly do not limit how much compute you can consume, they just auto scale it and then hit you with a big bill. I'm still relatively new to this so I do not want to end up like those students who accidentally setup some rogue EC2 instance that balloons to tens of thousands of dollars. I simply want a cloud provider where you prepay how much compute you want to use, and if you hit your prepaid limit, it just shuts down, no going into the red.

Or given this small setup, would it make more sense to not bother with the cloud at all, and spin up my own local server on raspberry pi ? Is all of the port forwarding, setup etc. significantly more complex than a cloud provider?

https://redd.it/1iuekh8
@r_devops

Should I use Terraform, AWS CDK, or bash scripts with aws cli???

What are your thoughts? Also, doesn't need to be only for aws, interested in hearing opinions from people working with gcp and azure as well, and comparing those apis with Terraform.

https://redd.it/1iue7u8
@r_devops

Hyperping vs. Better Stack vs. OneUptime for observability

Which one is better? Pricing is not the problem.

I am specifically interested in synthetic monitoring with playwright.

https://redd.it/1iugfm8
@r_devops

Community Powered Cloud based on TEEs

Since AMD SEV-SNP is now fairly easy to integrate on Linux, I believe that cloud will slowly start to move away from big centralized platforms. In order to start working with SNP, you need some Rust experience and I suggest starting with virtee: https://virtee.io/

AMD SEV-SNP is focused on creating Virtual Machines. VirTEE offers SNP integration for QEMU, and the old technology (SEV) is also integrated fine with libvirtd. Intel offers alternative technologies: Intel SGX (that offers containers, and that is older and more mature in terms of frameworks and implementations) and intel TDX (that offers VMs and is very new).

We made the decision to go down this path for our cloud start-up. We just created a testnet and are looking for feedback. If you would like to know more, I wrote a blogpost about it: https://medium.com/detee-network/so-we-have-a-testnet-now-2950de897ec6

https://redd.it/1iug77w
@r_devops