How future proof is DevOps?

I am sure a lot of people ask this question, but I haven’t found a backed reason as to why it’s good to learn it.
I’m a student who is interested in pursuing a career in DevOps, I barely have any experience yet except for mainly FE and BE basics with some DB knowledge.
In general how much is the demand for DevOps engineers and are the salaries good for Europe?

https://redd.it/1k5w8t3
@r_devops

Top devsecops interview questions

I just completed a devsecops course, ECDE to be precise, and I started getting multiple call when I update my resume. I have crack 3 interview and this is what I found they are mostly asking for.

* Can you discuss your experience with implementing and managing CI/CD pipelines?
* What are some common challenges you have encountered when integrating DevOps practices within an organization, and how did you overcome them?
* Describe your experience with containerization technologies such as Docker and orchestration tools like Kubernetes.
* Have you worked with any configuration management tools such as Ansible, Chef, or Puppet? Can you explain how you have used them in your previous projects?
* Can you discuss your experience with infrastructure-as-code (IaC) tools like Terraform or CloudFormation?
* How do you ensure high availability and scalability in a cloud-based infrastructure? What strategies or tools have you used?
* How do you ensure secure coding practices within a DevOps environment? Can you provide examples of security measures you have implemented?
* Have you worked with vulnerability scanning tools or security testing frameworks in a DevSecOps context? Can you discuss your experience and how they contribute to overall software security?
* Describe a time when you identified and resolved a critical security incident within a DevSecOps environment. What steps did you take, and what was the outcome?

https://redd.it/1k5ww79
@r_devops

Deeply curated database of 750+ well-funded, Remote-friendly startups + jobs

No, this isn't another scraped spreadsheet or pay-to-play directory. It's an open, manually curated database of well-funded startups building interesting things. Hard to find through all the LinkedIn/Twitter noise. And yes, I know startups aren't for everyone, but these are hopefully the better ones. Let me know what you think and hopefully it's helpful to find some interesting opportunities this year: hhttps://startups.gallery/

https://redd.it/1k5z3lv
@r_devops

Pull my head out of my arse on ai agents

I've been using github copilot for awhile. It's ok. My company is pushing AI pretty hard (like everyone else) and we all have a cursor licenses. Again, it's ok. I like the model as something to rubber ducky with and the agent mode to browse through files in an application to answer questions is neat. However, it seems like the industry is pushing more and more towards agentic implementations. Internally, I'm struggling with the idea. I'm in my mid 30s and have been at this for awhile. So this isn't "get off my lawn", but "how can i make something that I won't hate myself for in 6 months".

1) I was watching a video this morning /w bedrock and someone creating a customer service agent to process returns. The ideas are simple enough: model, couple lambdas, and some simple instructions. However, what's to keep the model from hallucinating at any point either to the lambda payload or the customer? We don't really have much control over the outputs. Sure, I could force feed them back in, but again I'm sending more and more requests to a black box. My underlying concern is when I or anyone else pay for a service, we expect that service and want it to be consistent. It seems dangerous to me that we're moving *stuff* out of known happy paths and into a magic box.

2) I've been reading some interesting details on model posioning. At the moment, it's typically by nation states who want to push certain view points and not underlying logic manipulation. However, the concern is still there. I can have code that doesn't change or I can ship requests off to a 3rd party model that could vastly change over time because the data being trained on has changed.

3) Just...why? While there may or may not be a cost savings from human labor (i have no idea i haven't done the math myself), it costs so much more to run a model perpetually than it would to have a web form that links back to the same lambdas.


I have a couple more, but am i wrong in thinking that while the models are neat, it doesn't seem like a great idea?

Regardless, announcements like shopify where they won't hire folks unless they prove it can't be done with AI are rampant and I have to adjust to die, but I don't want to go into that future with my eyes half closed from marketing gimmicks.

https://redd.it/1k6093x
@r_devops

Cloud vs Self-Hosted Logging

I'm working on a personal project (SaaS, not launched yet) and need to set up logging.

I'm considering two options:

1. Self-hosting a logging stack like ELK or EFK
2. Free/low-cost cloud-based logging service. I've seen that New Relic has a free tier with a 100GB per month ingest limit, which seems promising. I'm open to other alternatives as well (didn't do much research here).

What would you recommend and why?

https://redd.it/1k61r56
@r_devops

Built a Custom Kubernetes Operator to Deploy a Simple Resume Web Server Using CRDs

Hey folks,

This is my small attempt at learning how to build a custom Kubernetes operator using Kubebuilder. In this project, I created a custom resource called Resume, where you can define experiences, projects, and more. The operator watches this resource and automatically builds a resume website based on the provided data.
https://github.com/JOSHUAJEBARAJ/resume-operator/tree/main

https://redd.it/1k62tgz
@r_devops

There is a possibility that my org may implement DevOps practices…

Hey all!

I made a post here the other day asking about Terraform and CaC tools.

I was given great advice and useful information.

I wanted to reach out and actually provide an update regarding a possible opportunity and possible changes.

The org I work for is a global enterprise. We are a Windows/ Azure org. Our infrastructure is on-premise and in the cloud. I believe we recently moved away from physical servers and now host them using Azure VMs. Not sure if they use Linux or Windows servers though. I’m not that informed.

A year ago, I reached out to the cloud operations lead for the Americas (CAN, USA, LATAM). He told me to study Azure and I may be able to join the team someday. Well, I studied but they ended up hiring someone a bit more experienced. I cannot say I blame them. They were building up that team and needed more experienced people. Instead of holding a grudge, I reached out to the new hire and learned a lot of from him. He actually falls under my region of support so it’s normal that we communicate. Anyways, I eventually asked him about infrastructure as code and how much we used and what tools we used. Currently, the team doesn’t practice DevOps methodology so he didn’t speak much about. Instead, he referred me to the cloud operations lead. I reached out to the lead this morning and randomly just asked him if they were going to hire people once the hiring freeze was over. To my surprise, they are going to hire some people for junior opportunities. This time though, his advice on what to learn was a bit different than before. He advised that I study IaC (Azure native tools such as Bicep, and ARM) and CI/CD pipelines. It seems that my company may start practicing DevOps. Or at least, that is my takeaway.

I’m not sure how much time I have but I was able to get a voucher from MS. AZ-204 is one of the exams I can take for free using this voucher. I’m going to study this and then study AZ-104.

Wish me luck all! This may be my way in! I’m hopeful and excited!

https://redd.it/1k649ri
@r_devops

Devops/SRE AI agents

Has anyone successfully integrated any AI agents or models in their workflows or processes? I am thinking anything from deployment augmentation with AI to incidents management.

-JS

https://redd.it/1k69d11
@r_devops

Bad interview asking for reference from 10 years ago

I just wrapped up an interview, it started out well until the interviewer asked if I could provide references for two of the companies that I worked for in the past. One of those companies was from over 10 years ago, so I politely asked him if he meant another company with a similar name. He said no, he meant the company from 10 years ago. At this point I have a confused look on my face and before I could even tell him that I could provide a reference from that company (even though I thought it was strange given the time and that it wasn't a DevOps role), he goes 'Yeah the company's on your resume isn't it? You work there didn't you?'. At this point I'm all sorts of confused and flustered. I tell him yes I did work at the company and before I can say anything else he says 'you don't keep in touch with people'. I tried to explain that I haven't really kept in touch with anybody from my time there and that I've been out of the local market for a while (don't know why I mentioned that and I regret it now), but I could provide my manager's information. He then goes on to ask me what's wrong with the local market and as I'm answering his question abd talking about how bad the local market is, I'm thinking why am I even talking about this right now? We end up moving on to technical questions, things like ' how does DNS work?', ' how does a CDN work?', ' how does terraform work?', etc. but at this point I'm so flustered and confused about our 10-year-old reference argument that I struggled to answer these basic questions. I honestly don't even understand how a reference from 10 plus years ago and a different role would even be helpful. People change a lot in 10 years and most people don't clearly remember 10 years ago.

Has anyone else been asked for reference 10 plus years ago?

https://redd.it/1k6ai4w
@r_devops

Managing Deployments of gitrepos to servers

I am slowly getting into to devops, however the plethora of tools which all seem to market themselves as the solution for everything it's pretty hard to figure out which is the right way to go. I hope this subreddits experience can guide me in the right direction.

I am managing a variety of services for multiple clients. Each client has one or more vps instances containing multiple services, all running as a docker compose project. Each service has its own git repo, some are client specific (websites) and some are general and reusable (reverse-proxies, paperless, etc.).

I'm now trying to figure out what the best way to approach deployments and updates would be.

My ideal scenario would be a tool which would allow me to:
- Configure which repo (and version) should deploy to which server.
- Execute a workflow/push the repo using ssh-access from a secrets' manager.
- Monitor whether it is successful or not.

My only requirement is to self-host it.

Would gitea or jenkins be the best way to approach this? Thanks for any insights.

https://redd.it/1k6c58t
@r_devops

Is devops relatively hard field to get into as new grad?

How did you get your first DevOps job?

https://redd.it/1k6bwvh
@r_devops

Can’t get UTM data from HTML forms

I'm creating an HTML form to embed in Framer (so that I can get around the limitations that Framer places on form response submissions). I've already managed to create the forms and send the information to my webhook.

The only problem is that I can't capture the page's UTMs via this form... Is this the best solution? Has anyone who knows about Framer ever experienced this?

https://redd.it/1k6artb
@r_devops

Have you built QA/Testing pipelines?

In my experience I built CI/CD pipelines for Dev, Stagging, Prod environments but I never really built a pipeline that did automated testing. It makes to not have it in the prod pipeline. But I’m curious, if you guys have built such pipelines. If yes, what can you share about it? How did it integrate with your CI/CD overall?

Edit: I only have 1.5 years of experience in DevOps and it was my first fulltime job

https://redd.it/1k6ijz2
@r_devops

Tired of setting up the same pipelines? I'm building a CLI that deploys projects with natural language.

Starting a new service usually means hours of boilerplate: creating GitHub repos, setting up tests, Docker images, CD pipelines… What if you could just describe what you want?

I’m building 88tool, a terminal CLI that uses AI agents and LangChain to plan and execute full deployment pipelines.

It supports Go, Python, Java, etc., and connects to GitHub, AWS, Vercel, and more.
It’s not just generating code — it runs it.

Would love to hear from fellow devs who struggle with CI/CD fatigue.

https://datatricks.medium.com/building-in-public-from-terminal-to-deployment-with-ai-driven-ci-cd-fca220a63c58

https://redd.it/1k6kflk
@r_devops

pfsense ipsec tunnel aws issue

I know i can connect to two vpc via peer connection or transit but i need to get myself familiar with pfsense.

Current setup.

vpc1 (172.31.0.0/16)

pfsense1 (172.31.0.100) with public ip address
test1-ec2(172.31.0.101) no public ip address

vpc2(10.0.0.0/16)

pfsense (10.0.0.100) with public ip address
test2-ec2(10.0.0.101) no public ip address

1. Setup ipsec tunnel IKEv1 between the two pfsense. Both phase 1 and phase2 connection establish.
2. Both pfsense instance can ping each other (icmp) from their private ip address. So 172.31.0.100 can ping 10.0.0.100 without problem.
3. The route table attach to the subnet on vpc1 is routing traffic of 10.0.0.0/16 to the pfsense1 eni while the vpc2 route table routes traffic to 172.31.0.0/16 to the pfsense2 eni.
4. configured the firewall -> rules -> ipsec to have source and destination respectively. so for pfsense1 source is 172.31.0.0/16 to destination 10.0.0.0/16 all port and gateway. Vice verse for pfsense2
5. firewall -> nat -> outbound set to Automatic outbound NAT rule generation. (IPsec passthrough included)
6. the security group attached to both ec2 have icmp enable to 0.0.0.0/0

However test1-ec2 cannot ping test2-ec2 nor pfsense2 vice versa, `traceroute` gives me nothing but `* * *`

What am i missing here?

https://redd.it/1k6k5vg
@r_devops

How do you learn new setup and then impart the knowledge to others in team?

This is a slightly different kind of question.

We're using EKS with KEDA to run agents in our Azure DevOps pipelines. This entire setup is deployed using Azure DevOps pipelines (executed via Azure agents) along with Helm, ArgoCD, and Terragrunt.

The challenge is that this setup and pipeline were created by someone who is no longer part of the team. I’ve now been assigned the task of understanding how everything works and then sharing that knowledge with the rest of the team. We have created a user story for this task :D

The issue is that none of us has much experience with Kubernetes, Helm, ArgoCD, or Terragrunt. So my question is: how would you approach a situation like this? If someone could break down their process for handling such scenarios, that would be really helpful.

My main concern is figuring out the most effective and efficient way to learn the setup on my own and then transfer the knowledge to my teammates once I’ve understood the setup myself.

Thanks

https://redd.it/1k6ozjy
@r_devops

how to pass env variables to docker container when using github actions

how to pass env variables to docker container when using github actions to build image and running the container on linux virtual machine

currently i am doing this -

docker run -d --name movieapiapp_container \

-p 6000:80 \

-e ConnectionStrings__DefaultConnection="${{ secrets.DB_CONNECTION_STRING }}" \

-e Jwt__Key="${{ secrets.JWT_SECRET_KEY }}" \

-e Jwt__Issuer="web.url\

-e Jwt__Audience="web.url\

-e ApiKeyOmDb="${{ secrets.OMDB_API_KEY }}" \

-e GEMINI_API_KEY="${{ secrets.GEMINI_API_KEY }}" \

-e Google__Client_Id="${{ secrets.GOOGLE_CLIENT_ID }}" \

-e Google__Client_Secret="${{ secrets.GOOGLE_CLIENT_SECRET }}" \

-e ASPNETCORE_URLS=https://+:80 \

is this correct or is there any better way to pass these env variables ?

https://redd.it/1k6q5m3
@r_devops

First AWS cert to go for ?

I’m a software development engineer with 3 years of backend experience and I’m looking to transition into cloud computing, specifically with AWS. Which AWS certification would be the most suitable to start with?


https://redd.it/1k6t5q4
@r_devops

What happed to the DevOps Paradox podcast?

The DevOps Paradox podcast is my favorite and they haven't done a show since February.

Does anyone know why??

https://redd.it/1k6ujiv
@r_devops

Exploring Serverless Stack Architecture – How Do You Manage Environments & Security?

Hey folks,
I’m experimenting with a serverless stack on AWS using S3 + CloudFront for static hosting, API Gateway + Lambda for backend, DynamoDB for data, and Cognito for auth.

It’s been great for learning, and I’m thinking ahead about how to scale and manage this more professionally.

Curious to hear from others:

* How do you structure environments (dev/staging/prod)? Separate accounts, or manage via IaC/tagging?
* Best practices for securing this kind of stack — IAM roles, access boundaries, etc.?
* Any underrated tools or AWS services that help you keep things maintainable and cost-effective?

Appreciate any insight — always looking to learn from real-world setups. Happy to share my setup later once it’s more polished.

https://redd.it/1k6sux8
@r_devops

Best Practices for Horizontally Scaling a Dockerized Backend on a VM

I need advice on scaling a Dockerized backend application hosted on a Google Compute Engine (GCE) VM.

# Current Setup:

* Backend runs in Docker containers on a single GCE VM.
* Nginx is installed on the **same VM** to route requests to the backend.
* Monitoring via Prometheus/Grafana shows backend CPU usage spiking to **200%**, indicating severe resource contention.

# Proposed Solution and Questions:

1. **Horizontal Scaling Within the Same VM**:
* Is adding more backend containers to the same VM a viable approach? Since the VM’s CPU is already saturated, won’t this exacerbate resource contention?
* If traffic grows further, would scaling require adding more VMs regardless?
2. **Nginx Placement**:
* Should Nginx be decoupled from the backend VM to avoid resource competition (e.g., moving it to a dedicated VM or managed load balancer)?
3. **Alternative Strategies**:
* How would you architect this system for scalability?

https://redd.it/1k6x7tp
@r_devops