Seeking Some Words of Wisdom

Hi all,

I’m currently working as a Platform Engineer at a large multinational company, but my journey here has been anything but straightforward. I started my career 11 years ago as a .NET developer. After a few years, I began feeling stagnant and found myself drawn to the world of cloud technologies. Driven by this passion, I started teaching myself everything I could from online tutorials and guides, determined to gain the necessary skills in platform engineering.

About 3-4 years ago, I took the leap and fully transitioned into the Platform Engineering space, and, I’m happy with that decision.
However, I’m constantly reminded of just how fast the world of DevOps evolves—especially with the rise of GenAI, MLOps, and other emerging technologies. It’s exciting, but also overwhelming at times.

No matter how much I learn or how many projects I work on, I can’t shake the feeling that it’s never enough. I struggle with the question of whether I’m truly “qualified” to call myself a Platform Engineer. I don’t hold any formal Kubernetes or cloud certifications, but I’ve gained hands-on experience working with these technologies. Still, the lingering doubt remains—how much is enough?

I find myself feeling uncertain about areas like networking and Linux, especially since I transitioned from a purely Windows-focused background. This sense of not knowing enough sometimes makes me question my place in this field.

I’m hoping to hear from others who may have faced similar feelings or have advice on how to navigate these challenges. How do you balance continuous learning with feeling confident in what you already know? How do you define “enough” in a field that never stops changing?


https://redd.it/1g7684l
@r_devops

Automate Deployment config changes?

There is something I have always been wondering about how to best solve this. The problem:

Deployments in Kubernetes cluster based on helm and ArgoCD. Now most things can be automated quite easily with this setup but what always seems to become troublesome in bigger projects are changes to configmaps and secrets especially staging these when they are environment specific.

Current setup:

Developers try to document all required changes and set values in a secret store that is referenced. This however still requires a lot of effort before deployments to change some environment variables in helm charts and secret references etc.

Is there a setup to fully automate this easily? We have a ton of different staging environments >25...


Edit: All generic environment variables and configmaps get baked into the helm base charts/images already

https://redd.it/1g76zcp
@r_devops

Seeking Career Progression Advice

Hey everyone,

I'm looking for some guidance on my career progression and would really appreciate your insights.

I started off as a DevOps engineer focused on containers, cicd, automation scripts, etc. Over time, my career led me more towards AWS, and I ended up obtaining several AWS certifications (8). This compounded my interest and skills in AWS, and I guess you could say I'm a cloud engineer now. I still support applications through cicd and containerizing, etc, but that's not the highlight of my skills.

Lately, I've been feeling like I've been stuck in the same role or rut for too long. I considered moving towards DevSecOps or application security, but that seems to be along the same track I started off in (containers and pipelines).

I'm curious about what career progression paths I can explore now. Given my background, what roles or areas might be a good fit? Has anyone else been in a similar situation and made a transition that revitalized their career?

Thanks in advance for sharing your thoughts

https://redd.it/1g7a9ep
@r_devops

How to authenticate paid software?

Context:
I have a freemiun app, most features relies on a local LAN server, except for a proxy server and in app features for the admin of said server.

What I've been thinking..

Method 1
- Being online most of the time to ensure the user is using a valid key, but this creates a conflict with the core of the app (minimal use of the network)

Method 2
- Ship the app with a public key to validate the user key, this key has encrypted data of the user and expiration date, this is stored on the client's device. Upon key expiration it's invalidated and removed from the device. The user has to pay again. (This one only relies on the network once per key activation)

What other methods can you think of for this context?

https://redd.it/1g7hg5k
@r_devops

Terraform Test or Terratest

We currently don't have any sort of unit tests for our modules and I was checking into both of these tools. They both have their pros and cons, but I wasn't sure what people thought about the pros and cons of each at this point. I picked up enough Go to play around with Terratest and I do like how easy it is to import a module and do basic things that I'd want to do. Of course the native Terraform testing framework uses HCL so that's a big plus, but I don't feel like there are many examples out there comparatively and that it would be harder to achieve certain tests not having the flexibility of a programming language like Go. Curious to hear what you think.

https://redd.it/1g7iyeh
@r_devops

Is it good to learn DevOps through openstack

Hi folks
I have 6 nodes and i wanna learn DevOps through openstack
What are you opinions about this

https://redd.it/1g7i7iw
@r_devops

Suggestions on what to improve on in order to take bigger rolesband advance career.

I have been at my current place for 6 years and it's been a team of 3 managing a multimillion infrastructure spend. I came on to migrate us from on prem to azure and I had full control of what we use. I recruited another engineer I knew at my last job and we have run a really stable environment over the last few years with the third guy that was hired from within the company.

We use terraform, cinc, PowerShell, and jenkins to manage all infrastructure and automation. I am extremely strong in azure and PowerShell. I would say fairly strong in the other two, and I have a little experience with some other things..

We are a msft shop and everything is running on VMs but we've got some things in Linux and we're integrated in some azure services like azure files, redis, blob, and networking services like virtual wan.

We were moved to GCP after we got bought and its ok, but I'm getting bored. I like to do infrastructure architecture and environment build outs but its just maintaining a lot of it now.

Are there industry trends/new technologies that I should be learning about to further my career for other possible jobs? I was thinking learn some more on containerization but I'm not sure. I've done cloud infrastructure and automation for about 10 years now.

Any input is appreciated.

https://redd.it/1g7mwib
@r_devops

Memory waste and cold start

Drowning in config files and CloudWatch logs. Two issues I was thinking about sharing with you and gathering your thoughts:
1. Memory Waste: A study found 95% of serverless functions use <10% of allocated memory. How are you optimizing this in your CI/CD pipelines?
2. Cold Starts: Balancing performance vs. cost is tricky. What's your strategy for managing this, especially for low-traffic functions?

I'm curious:
* What tools or custom scripts are you using for serverless optimization?
* How are you handling observability and cost management for serverless in your DevOps workflows?
* For those at scale: Any tips on maintaining CI/CD speed and configuration management as your serverless footprint grows?

https://redd.it/1g7ubcm
@r_devops

I can’t write the right vagrantfile for hyperv

I want to learn ansible but i can't start because i couldn't make the vagrantfile to run well on hypery (i can't use virtualbox) so if someone can help me adjust a vagrantfile so it can run on hypery with the same vms
Please send me a message if you can help and i will really appreciate it

https://redd.it/1g7vnr7
@r_devops

shared developer setup for macOS and windows

Hello,

recently we have finally been able to use MacBooks in our company, i.e. also for our customer projects.

Previously, development was done purely on Windows laptops and there was a setup with SEU-as-Code that created an isolated environment with IntelliJ, Git for Windows, VirtualBox for a Devbox and SQLPlus.

The environment was then assembled as required using configs, start and env scripts.

However, the whole setup is very old and should have been modernized a long time ago anyway, as SEU-as-code has not been developed further for years. Now they want a new setup and support for both operating systems

The current statement from our leads is that the costs for the modernization will increase by 50% if we want to support Mac and that they would like to use separate setups.

However, I find it hard to imagine that we are the first to run into this issue and that there is no practical solution, especially if the setup is developed from scratch and no technical debt has to be taken into account. Also, both leads were not fans of the Mac launch before, so I'm not sure if this is a purely objective opinion.

The question here in the round whether someone already has experience with this and perhaps has something like this in use in their own company.

Ideas, tools, anything would help, as I'm currently worried that our bosses might scupper the whole issue of Macs because of the statement.

Or does anyone know a better subreddit for the question if I'm wrong here, I haven't found one so I'm asking here in the hope it's OK.

https://redd.it/1g7x9vk
@r_devops

I wanna build my own cloud

Hi folks
I wanna build my own cloud that i wish to extend
Which approach do i need to go with and why

https://redd.it/1g83oes
@r_devops

Buildfarm on Kubernetes

Anyone running production-grade Bazel Buildfarm (I guess it's just Buildfarm now) setup using the official helm chart with Linux, Windows and macOS workers? Just curious about experience with it, especially autoscaling, CAS management, overall configuration etc.

I have a setup with all that, but without autoscaling (because when Linux workers scale down buildfarm still looks for what was in their CAS for some reason and builds fail when they don't find something that was cached - instead of just rebuilding it, for some reason). Windows workers performance is kinda dogshit, but that may be about our custom toolchains - or just Windows being Windows keeping that CPU usage under 40% pretty much always.

As we're on AWS / EKS, we're also thinking about moving CAS to S3. Anyone here had something like that?

https://redd.it/1g84lz5
@r_devops

Software Engineer Jobs Report 10/16: Every week I spend hours scraping the internet for recently posted software engineer jobs. I hand pick the best ones, put them in a list, and share them to help your job search. Here is this weeks spreadsheet. 260+ roles USA and aboard. Devops roles included.

Hey friends, every week I search the internet for software engineer jobs that have been recently posted on a company's career page. I collect the jobs, put them in a spreadsheet, and share them with anyone whose looking for their next role. All for free.

I hand pick the ones I know are good roles, with market salaries, and no glaring flags (ex: I generally only put roles with posted salary bands). Though its not easy to tell if the roles require leetcode or not. I want to figure out how to get the information in the future.

The data is sourced by my own web scraping bots, paid sources, free sources, VC sites, and the typical job board sites. I spend an ungodly amount on the web so you don't have too!

About me, I am a senior software engineer with a decade of work history, and ample job searching experience to know that its a long game and its a numbers game.

If there are other roles you'd like to see, let me know in the comments.

To get the nicely formatted spreadsheet, click here.

If you want to read my write up, click here.

if you want to get these in an email, click here.

Cheers!

https://redd.it/1g85r21
@r_devops

automate VPS server migration - what do you use?

im tired of doing it by hand - so I got ansible setup and can provision new hosts quickly.

But what do you do for data/application migrations old vps to new vps?
is it possible to automate that as well? using what tool? just rsync or is there anything better?

https://redd.it/1g85z0g
@r_devops

Simple helm charts to learn from

We're going through a k8s transition at work, and I've been through all the KodeKloud learning paths and read a lot of tutorials online about helm charts, but I've found either:

1. They're just tutorials that cover the concepts, and so they're very easy to understand.
2. Looking through random real-world charts like bitbami charts are super complicated and I have a lot of trouble understanding them.

I've tried googling for "simple helm charts" and stuff like that without success.

Does anyone have any hints on what I should do next to learn more so I can get closer to understanding the complicated bitbami and other complicated charts?

Thanks!

https://redd.it/1g8c2xh
@r_devops

Got Github actions running 2x-10x Faster with an overclocked self hosted actions runner. Benched against 3 CPU types (i9, EPYC, Threadripper) and Github's runners - How to guide

Keen for ways to make this go even faster, let me know if you've tried anything like this:

https://words.strongcompute.com/p/maximising-github-actions-efficiency

Other hardware things we're doing to speed up dev:
* M4s as soon as they drop, have a couple M3s already and they're compiling about 2x faster than the M1s.

(Most of our dev is in Elixir)

https://redd.it/1g8c9uh
@r_devops

Should Data Quality Be Managed in DevOps Sprints?

My experience with managing issues with data isn’t as simple as throwing them into a sprint, following the ‘tasks’, then closing out the user story at the end of the sprint. It simply does not work because discovery is a bitch. Furthermore, the emphasis is not on pipelines or the nature how of how data behaves that constitutes “quality”. Instead, I find that managing data issues involve a lot more stakeholder managing, identifying and reducing bottlenecks of trying to fit data into DevOps, and as a result blame from unaware or even arguably incompetent management that hide behind scrum because they do not have one clue about actual problem solving.

I cannot be the only one. Every suggestion I have made to prove that either our approach is inefficient, lacks necessary clarity, or just the wrong tool from the job gets met with some form of perceived dismissiveness on my end. Maybe I’m being too kind. Perhaps my manager is just plain inept at understanding how some things work and loves to just point fingers. I get the sense that they simply cannot believe I could be right. What’s apparent is that they never have anything remotely to contribute and it’s easy to tell.

Maybe I am ranting but I think DataOps could be a candidate for a proposed solution. Until, I’ll keep hammering this squared peg into this round hole.

Does anyone else have experience with this? Any suggestions for a different or modified approach with DevOps? I can’t help but to feel overwhelmed. All my ideas are mostly passed over if they mean changing the system. The others ones just get appropriated.

https://redd.it/1g8h6fn
@r_devops

Linux engineer with some devops experience, and I was recently laid off. I want to improve my devops skills but not sure where to focus.

So I'm a recently laid-off engineer, but I was working in a devops role. My role was kind of hybrid devops / linux engineer / cloud engineer. I have a ton of experience in system engineering and AWS, but as I start looking more for work I'm a bit unsure where to shore up my skills.

Things I definitely know I need to do are:

* Learn a scripting language - python or go seem to be most sought after right now. If I go the route of python I'll probably work off of Automating the Easy Stuff with Python.
* Learn more about CI/CD - my CI/CD knowledge is extremely limited, and I'm not sure where to start. Any recommendations accepted.
* Learning more about containerization. Again, recommendations for materials are greatly appreciated.

I already know a good amount about Kubernetes as it's inner workings, having been the go-to fixer for a large kubernetes cluster, so I am prepped there but it does seem like many companies are moving away from Kubernetes now. Not sure if I should be focusing my energies elsewhere.

https://redd.it/1g8jgfb
@r_devops

Discussing Challenges and Solutions in Automating PHP Application Deployments

Hello fellow DevOps practitioners,

I've recently been working on automating PHP application deployments, focusing on integrating GitHub Actions, AWS CloudFormation, and Ansible. I'd like to share some insights and open up a discussion about the challenges and solutions in this space.

Some key points I've encountered:

1. Balancing flexibility and standardization in CI/CD pipelines for various PHP frameworks (WordPress, Drupal, custom apps).
2. Managing infrastructure as code effectively with CloudFormation, especially for multi-environment setups.
3. Ensuring idempotency and consistency in server configurations using Ansible.
4. Implementing secure practices throughout the deployment process.

I've found that this combination of tools can significantly reduce deployment times and human errors, but it comes with its own set of challenges.

What has been your experience with automating PHP deployments? Have you encountered similar issues or found innovative solutions? I'm particularly interested in hearing about:

Strategies for handling database migrations safely
Approaches to blue-green deployments for PHP apps
Techniques for managing environment-specific configurations
Best practices for secret management in PHP deployment pipelines

https://redd.it/1g8krt6
@r_devops

Looking for DevOps mentor/guide

Hello. As the title says I'm looking for someone who is experienced in devops and willing to guide others and help in clearing any doubts in their free time. I was working as a devops engineer myself but the project i was working in had very little to do with devops and most of the work I had was just in linux servers and monitoring. I've nothing devops related to show on my resume. There was no proper planning in our project during deployments as well, no processes followed. This has made it extremely hard for me to find a job elsewhere as my exposure to the tools and devops methods are close to nothing. If there is any existing discord group or anything please let me know. I can go ahead and create one otherwise if anybody is willing to join and guide/ clear doubts and help in gaining actual corporate project knowledge on processes and methods followed.

https://redd.it/1g8k3pv
@r_devops

Please review my resume

I posted it last time and got some suggestions and redid the resume,

please review it and let me know any changes are needed.

here is the links https://imgur.com/a/BOt2CTC


and in my current position i already completed 2 years i spited it because it got an promotion.

https://redd.it/1g8mc3c
@r_devops