Why is DevOps still such a fragmented, exhausting (and ofc costly) mess in 2025?

I have been thinking about this for quite sometime and thought of getting your thoughts. I feel like DevOps was supposed to make life easier for developers, but honestly, it still feels like an endless headache. Every year, there’s a new tool, a new “best practice,” and a new wave of people claiming they have finally cracked the DevOps code… yet here we are, still dealing with the same mess, just with fancier buzzwords.

A few things I keep running into over the years that I have worked with different projects:

1. The never-ending toolchain puzzle – Every company I have worked with has a bloated DevOps stack. Terraform, Kubernetes, Jenkins, ArgoCD, GitHub Actions, Helm, Spinnaker—you name it. It’s like every tool fixes one thing but breaks another, and somehow, the entire setup is still fragile as hell. Instead of simplifying DevOps, we’re just stacking more complexity on top of complexity.
2. Burnout is real – I don’t know a single DevOps engineer who isn’t constantly tired. Between keeping up with cloud providers, maintaining brittle pipelines, dealing with security updates, and being on-call for random failures at 2 AM, it’s no surprise people are leaving the field. We were supposed to be automating things, not babysitting them 24/7.
3. Automation is a lie – Every new trend is supposed to “automate everything,” but in reality, we just end up automating a different kind of chaos and it becomes totally fragmented. IaC is great until Terraform state breaks and you’re in hell. GitOps is cool until you realize drift is inevitable. Pipelines are supposed to “just work,” yet half the time, debugging a failed deploy feels like solving a murder mystery with no clues.

And here’s the kicker: this mess is costing companies millions. There’s actual research backing this up:

The [2024 State of DevOps Report by Puppet](https://www.puppet.com/blog/state-devops-report-2024) talks about how DevOps is still in a weird transition phase, with more companies shifting towards platform engineering but still struggling with inefficiencies.
The **DORA 2024 Accelerate State of DevOps Report** highlights that while elite teams are getting better, the majority are still facing the same bottlenecks we’ve seen for years.

So, I gotta ask—what’s the real solution here? Has anyone actually figured out how to do DevOps without it turning into a soul-sucking nightmare? Or are we all just stuck in an infinite loop of new tools, more YAML, and never-ending on-call rotations?

Would love to hear how others are dealing with this. Maybe I’m just jaded, but damn, it feels like we should be further along by now.

https://redd.it/1i538r1
@r_devops

AI and the future of DevOps engineers

We've heard the news of massive layoffs in large FAANG companies for software developers, engineers etc. And with Mark Zuckerberg mentioning in a recent interview that more than 20% of junior devs are going to be replaced by AI; I'm curious to know what your outlook is for the future of DevOps engineers.

I appreciate that DevOps was originally supposed to be a philosophy instead of a job title, but how are you pivoting your careers, or not, with the advent of AI? Some of my friends are pivoting into cyber security, solutions architecture etc.

https://redd.it/1i54fag
@r_devops

Recommendations for DAM (Digital Asset Management) Tools/Products?

Hi Everyone,

Like many others - my company users a CDN service (CloudFlare in our case) to serve static files to our customers.

The CDN Origin in our case, is a simple IIS website, hosted on one of our on-prem servers - that serves up a folder filled with many static files organized (actually not really organized) in many folders and subfolders.

We've reached a critical point where a lot of people from a lot of different departments are accessing the folder and updating resources in it all the time (Developers, Designers, Marketing, Content).

We'd like to have some kind of control over the folder:

- Access control to specific paths in the folder (Who can access and modify what)
- Change review and approval before applying changes.
- History, with the ability to restore a file from older revisions.
- Easy to use by non-technical people (Designers, Marketing, Content)
- Supports Google (SAML2) SSO
- Not too expansive
- Easy to integrate

Unfortunately management has decided, disregarding the many arguments I presented against it - to use our GitHub enterprise for this, and had me create a repository with gigabytes of media files in it.

I am utilizing git-lfs in my solution, and have a GitHub action that runs "git pull" in the CDN directory from our GitHub Enterprise "CDN-QA" repository after each successful pull request merge to the master branch.

This is currently undergoing a pilot in our QA environment and is expected to go Prod soon.

I know there are MUCH BETTER solutions for this... Shelf products that are designed to be the solutions for our requirements.

I'm looking for recommendations based on your own personal experience:

Can you kindly recommend, some sort of Digital Asset Management system that you think we can utilize?
It should answer all of the above requirements and hopefully be easy enough to integrate and not that expensive?

Anything from ResourceSpace to Bynder or anything else that you think matches the described scenario?

Thank you kindly in advance!

https://redd.it/1i4ztj8
@r_devops

How do you handle the XY-123456 ticket is deployed to ZX-Stage question?

Hello!
Our projectmanagers and developers always ask XY-1234 ticket or feature is deployed in ZY stage? We are try to use gitops, but noone want to open the git to check it. Argocd web ui is also top much for them. We have Jira, gitlab, argocd, Google Chat.

There are any way I can display this information in Jira?

https://redd.it/1i56hj1
@r_devops

For all self-taught developers - what would you say is most difficult about building apps?

If you're building an app on your own, what do you think is a difficult thing about making an app on your own from scratch? I would assume that, being self-taught and solo, it would be difficult to properly understand, plan, and stick-to, one thing. An app must be planned out thoroughly, its timeline and scope and what not, and one must stick to developing it for multiple weeks or, more likely, months on end. And that, coupled with the fact that full stack is hard to build since you have to always understand how everything interacts between the backend and frontend. I'm doing some research to understand self-taught solo developers' experiences, any words would be appreciated. Thanks.

https://redd.it/1i59msb
@r_devops

DevOps Certification

Hello guys,

What are the top certifications for each category:
1. Automation and CI/CD
2. Containerization and Orchestration
3. Infrastructure as Code (IaC)
4. Cloud Platforms
5. Security (DevSecOps)

https://redd.it/1i5b3nr
@r_devops

Adding Macs to on-prem build agent pools

We have a few build jobs that build on arm64. So far I used buildx with the qemu bin tools and while it works it is slow, and even worse, some of our build jobs now fail. I am not quite sure what happens, but cargo stops to make any progress.

As we have our build servers on-premise, I was wondering how I could add native arm64 machines. I thought that maybe using Macs would be ideal, as it allows us to support building for Mac and Linux-arm64, but I am not so sure about OpSec for Mac Server.

Does anyone have experience with Macs as build agent for azure devops?

https://redd.it/1i59wgc
@r_devops

How do people have a good devops portfolio?

What kind of projects can you have that will show your skills in devops? I'm curious how you show your skills around deployment, monitoring and other stuff, especially when talking about a project with a cloud provider...

Any help would be greatly appreciated!

https://redd.it/1i5dnwd
@r_devops

Portfolio needed for FAANG devops role?

Hi everyone, I’m currently in my 3 YOE as a platform engineer at my company and prior to that have 3 YOE in network security. I want to start looking into getting a devops/devsecops role with a FAANG and I’m curious if I’d need some sort of portfolio with projects to even be considered at these companies. I’ve done a decent amount of work surrounding AWS services, proficient in Terraform, have written a large majority of our Sentinel policies (security policies for terraform runs), worked with automation, know the ins and outs of standard CI/CD pipeline implementation w/ Concourse and GH Actions, but it doesnt seem like it would be remotely enough to be considered.

Does anyone have any insight on if Im being unrealistic with my goal of joining a FAANG in the next 1-2 years? Should I be focusing more on other aspects? Would the portfolio/project route be a waste of time? Any information would be helpful!

https://redd.it/1i5g1cg
@r_devops

Hashicorp Vault - Transit Secret Engine - Decrypt Mechanism

While using decrypt action in the Transit Secret Engine, we do not have the option to choose which version of a particular key we can use to decrypt a Ciphertext.

Is it because the Decrypt action is done using only the corresponding version which was used to encrypt initially? We don't have a flag to mention which version to use for decrypt action for a key.

For example: when we do the below action, does it automatically use the version 2 of the "test" key to decrypt the ciphertext?

vault write -f transit/decrypt/test ciphertext="vault:v2:fRds/te23Ra2KnsL+Jomk6ZYA4PS8uv/bbyjM0LDiNKfWOdk61vi4rvFMcClANUPvOc="

Can we decrypt a ciphertext produced by version 2 of a key, using version 3 of the same key?(without rewrapping)

https://redd.it/1i5mj47
@r_devops

Pipelines with ArgoCD

I have to use Argo now and was used to push based gitops before. Before e.g. I used Gitlab Pipelines to install on Dev, then execute Tests and if Tests are successful, I have a button to manually promote to the next environment. Release to prod was then a manual pipeline.


So how do you handle processes like this with ArgoCD? I see there are tools like Kargo or Keptn or commercial tools Codefresh. So its seems I'm not the only person missing that on ArgoCD :-)

Can you guys tell how you handle such things and hint what to look out for?

Greetings

https://redd.it/1i5nfjh
@r_devops

How good is this for a person with some Devops experience

I have some prior Devops experience and want to further enhance my skills. I wanted to know how good of a resource is https://devopsroadmap.io if I want to get better at Devops?

https://redd.it/1i5nxbr
@r_devops

Is this too restrictive company policy?

Hi, I'm DevOps Engineer in local branch of big global corporation. My team is developing internal applications and I was hired to automate and simplify that process mostly by containerization. However every time I want to deploy some new infrastructure or tools I encounter multiple obstacles on every possible level.

On our work laptops we don't have classic admin rights but some pseudo-admin so changing any configuration, installing or uninstalling is almost impossible (for example I can't install Podman Desktop because wsl version is not supported and I can't update version because this is blocked by company).

Same goes to cloud infrastructure. We have many cloud service on Azure but we don't own any of that because it was outsourced to indian company. So can't change anything and have to create tickets to this company asking for that. We don't even see most of our resources in Azure Portal, so configuration like corresponding vnets or so are pretty unknown for us. For most services we have very limited rights to install or change anything. So anything deployed there is very fragile and needs constant monitoring, trouble shooting and work arounds.

Our network firewall is also very strict. Almost everything from open internet is blocked. Opening some port/ip even for well known sites, repositories (even jit for installing tools) is pain in the ass and almost impossible to get approval from info sec. And talking about infosec, they don't communicate with us, don't try to work with us. They just say no for almost everything. And when we ask for some rules, guidlines how we should work to met their demands and to ensure we can have tools and processes we want, they won't share anything. I have a feeling that the best for them would be if we worked in disconnected from outside world basement, so they won't do anything.

Is this normal? Or is this just my company?

https://redd.it/1i5pc9r
@r_devops

Newbie To Devops

Hi all,

I've been a 3rd line/sysadmin for 7 years now and we are moving to azure (NHS don't ask).

I've got myself a pluralsight sub and doing the azure labs and terraform and managed to work out the basics of deploying vm's using visual studio and even adding some extra modules in to some of the labs to further understand things. Even separating modules into the own files rather than using a flat main.tf.

My question to you gurus is, what is the main thing companies want these days?. Now i have also dabbled with ansible spinning up som vm's at home and doing some basic playbooks.

The problem is i see jobs advertised with puppet, chef, ansible, terraform ad infinitum. Seems overkill to me.

It's confusing as to what is widely used, and too be fair i have a partner and kids and i don't have the time to learn it all.

Thank you for your time and advice, It's much appreciated

https://redd.it/1i5nfjn
@r_devops

What do you recommend I should use for making my first CI/CD pipeline?

I was thinking about using jenkins? Is that still widely used?

https://redd.it/1i5sk7r
@r_devops

Made an app for Anime, Manga and Novel lovers written in flutter

Named AzyX https://github.com/reyyuuki/Azyx/releases

https://redd.it/1i5tynl
@r_devops

Cluster API to production: from Cluster API to GitOps with Argo CD and Kyverno

Hi everyone!
I just finished writing the second part of my Cluster API to production series: an article about configuring Argo CD to deploy applications on Cluster API tenant clusters with Kyverno.
The series bridges the gap between the Cluster API documentation and deploying production clusters.
The next part will be about configuring RBAC and service accounts for tenant clusters.

Even if you already have Argo CD set up, you might find the linked cluster Helm chart interesting.
The chart includes Argo CD applications configuring a telemetry exporter with OpenTelemetry Collector and FluentBit, centralized secret management and more.

https://redd.it/1i5ubjq
@r_devops

How to go from IT Support to Dev Ops (or even developer in general)

I have a portfolio. I have been coding in java for years and I feel i'm well versed. 4 years in IT, and my two related jobs are coding adjacent but not full on coding.

Problem is, all positions I see are Senior Java positions. I can't get junior experience, so how can I even land a Senior position. I'm working in a programming adjacent IT role(implementation and review of code whilst supporting IT tickets)

I have the desire and the knowledge to move forward, I have worked in a lot of different frameworks. I know the basics(Spring Boot, Docker, Git, Jenkins) and my ability to grasp new tools is almost instantaneous. BUT i need a direction of where to even go. Resources are so limited too after college.

Everyone seems to pretentious to help as well, they think it should just fall into your hands or that you should already know what to do. I don't know what to do. I straight up cannot land a new job at all so I want to develop my skills a bit to give me an edge

https://redd.it/1i5whgy
@r_devops

Building NixOS 24 Snapshots on Hetzner Cloud with Packer - My Learning Journey

Hey fellow DevOps engineers!

I've been wanting to try out NixOS for a while and finally took the plunge by setting up a proper build pipeline using Packer on Hetzner Cloud. I documented my experience in a blog post, hoping it might help others who are curious about the same stack.

What you'll find:
- Complete Packer configuration for building NixOS 24 snapshots
- The entire setup script including disk partitioning and NixOS configuration
- Real challenges I faced
- Bonus OpenTofu code for deploying servers from the snapshot

I'm definitely not a NixOS expert, and there might be better ways to do this. The configs are working but probably not optimal - I tried to document my thought process and include necessary explanations for each step.

If you've implemented something similar or have suggestions for improvements, I'd love to hear your approach. The main goal is to learn and share experiences with the community.

Link to blog post: https://developer-friendly.blog/blog/2025/01/20/packer-how-to-build-nixos-24-snapshot-on-hetzner-cloud/

https://redd.it/1i5v0ew
@r_devops

How much "Go" is needed to learn?

Im going through KodeKloud's devops path and I was surprised to find the "Go" language as one of the languages to learn. Im experienced with C++/JS so I was able to mostly "skim" through it. That being said how much (or I guess how deep) is go actually used in devops?

I know Kubernetes/Docker are written in it, but I'm not sure what all a devops would write in actual Go?

https://redd.it/1i60qso
@r_devops

Is Jenkins still popular enough to learn?

Pretending someone is JUST getting into devops, do you think learning Jenkins is worth the time investment vs other CI/CD tools (Gitlab/Github/Azure). Assuming you aren't already in a job that uses it.

FWIW: Im asking because KodeKloud has Jenkins on their devops path. Our company uses Gitlab so I figure it makes more sense to swap it out for their gitlab course.

https://redd.it/1i613zc
@r_devops