Configuration Variables

All my companies applications are configuration driven. At the moment we use Azure DevOps for CICD.

However, the library groups are awful and have no auditing and has grown out of hand. What are your methods for handling mass configuration? My idea was having a configuration repo which the applications can pull in and use.

If any advice, please share!

https://redd.it/1krpa76
@r_devops

Elasticsearch Labs

Hi all, can someone point me to the right direction so i can prepare my self for some interview that wants elasticsearch experience? platforms like kodekloud doesn't have labs for it unfortunately, thanks!

https://redd.it/1krqzg4
@r_devops

Ms teams chat bot

Hi guys,
We’re investigating if it’s possible to build a bot which communicates certain kubernetes actions from teams to a private aks cluster.

In our current situation we have a golang bot running in an azure container app which is connected to slack, this works perfect. The communication works via websocket which makes it quite easy to arrange this. But to my understanding ms teams does not support this. My knowledge with teams is quite basic so I’m kind of wondering if it’s even possible to rewrite this for teams.

Slack is being replaced by teams in my organisation (unfortunately) so hence the use case. I’m curious if someone has done this before and what their experience was like.

Thanks guys!

https://redd.it/1krr2qd
@r_devops

Vibe Coding is great until its not... How are you tackling this challenege personally or in your team?

I promise I’m not turning into a “back in my day” rant, but things just working is becoming rare.. only 3–4 years ago things where basic but bugs where rare to expierence. Yesterday, I was drafting an email in Gmail when suddenly the Send, BBC and Discard buttons just wouldn’t click, and entire lines of text duplicated themselves out of nowhere.

With the pace of software updates, shrinking dev cycles, and now this thing folks call “vibe coding,” it feels like on-call nightmares are staging a comeback.... only this time, nobody truly knows what they’re on call for 😭. Vibe coding can crank out features fast, but pushing it live without understanding its quirks (or owning up when something breaks) strikes me as downright reckless.

Back in the day, on-call meant a team of engineers who knew every corner of the codebase. Now? It feels like handing the keys to a car nobody’s test-driven. Sure, 100% unit test coverage looks great on paper, but it’s not the same as real world, black-box, user-centered validation.

So I’m curious: how are you folks testing or validating “vibe code” in your shops? Have you seen similar random tech gremlins, or is it just my luck? Let’s compare war stories—maybe there’s a better way to keep our digital lives from glitching into chaos.

https://redd.it/1krt8bo
@r_devops

I really hate working in tech but can't do anything else

I've been a Dev for over 20 years with some exposure to DevOps. I really hate everything about it - the people, the "culture", AI. I've gotten to the point where I can barely make myself go into work or even feign the slightest bit of interest / effort each day. Just doing the bare minimum to pass myself.

Anyone else feel like this? What are other potential careers where someone with a tech background can look to switch to? Literally anything would be better than this grey blandness.

https://redd.it/1krtx2h
@r_devops

I applied to 10,000 DevOps jobs

No, I’m not unemployed. And no, I’m not desperate.

I just wanted to stress test the DevOps job market in 2025, across regions and industries, using automation where DevOps shines: pipelines, scripts, and APIs.

Profile used:

Degree: MSc in CS , graduated with honors
Experience: 5 years, from infra-heavy Quantitative Analyst roles to full-on DevOps/SRE
Past employers: internal DevOps/data platforms team
Languages: Python, Bash, C++, basic Go
Tech stack:
Infra: Docker, Kubernetes, Terraform, Ansible, GitOps
CI/CD: GitHub Actions, GitLab CI, Jenkins
Cloud: AWS (EC2, S3, Lambda, EKS), some GCP
Monitoring/Logging: Prometheus, Grafana, ELK, Sentry
Security: IAM, Vault, secrets mgmt, RBAC
Extras: FastAPI for internal tooling, self-hosted runners, feature flag systems

# I didn’t just spray resumes.

I built a pipeline using:

**hiring.cafe** to scrape relevant jobs
OpenAI API to tailor resume & cover letter
**Laboro.co** to autoapply jobs

Every. Single. Application. Was unique, just not done by hand.

# Here’s what happened:

|Country|Applications|Human Interviews|AI Interviews|Assessments|Rejected|Ghosted|
|:-|:-|:-|:-|:-|:-|:-|
|USA|2,037|18|26|50|788|1,155|
|UK|1,503|9|18|52|500|924|
|Canada|1,025|7|14|30|392|582|
|Germany|935|5|12|16|358|544|
|France|822|4|10|22|254|532|
|India|748|3|9|18|298|420|
|Australia|630|3|10|12|243|362|
|Netherlands|536|4|8|12|206|306|
|Spain|416|3|5|10|178|220|
|Sweden|327|3|4|6|152|162|
|Remote|2,446|13|34|60|1,033|1,306|

# Key takeaways:

Remote DevOps roles are abundant, but have the highest ghosting rate.
AI screenings and coding tests are becoming more frequent, especially in enterprise/cloud-heavy roles.
European countries seem more balanced in feedback rate vs. applications sent.

# Why I did this:

I wanted to measure hiring friction, not just read Reddit threads.
I wanted real market signals, not opinions.
And let’s be real, building the automation to do this was the most DevOps part of the whole thing.

# Happy to answer:

How I built the application pipeline
How effective the resume tailoring was
How DevOps hiring differs by region or sector
Or just share scripts/templates if folks are curious

https://redd.it/1krwg92
@r_devops

What's the use of tools like Azure Key Vault, AWS Secrets Manager etc.?

Don't use .env files use Azure Key Vault!

To connect to AzureKV - you need to store client id/secret in .env which can be used to get those secrets.

If I have the .env file, I can get the secrets.

What I'm missing here? I don't understand...



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

How do you standardize dev environments across multiple teams and projects?

Curious how others are tackling this — especially in fast-moving teams with lots of microservices or side repos.

I keep running into the same friction:

* Inconsistent or outdated setup instructions
* Missing `.env.example` files
* Dockerfiles that break on fresh machines
* GitHub workflows that are unclear or undocumented
* Onboarding that relies on tribal knowledge or Slack archaeology

It becomes a game of “ping the last person who touched this,” and it doesn’t scale.

I've started working on a tool that reads the structure of a GitHub repo and **auto-generates all the key onboarding and setup files** — like `README`, `.env.example`, Dockerfile, GitHub Actions, etc.

Not pushing it here — just wondering:
**What strategies, templates or tools have you found effective to reduce this chaos?**
Are there standards in your team for onboarding-ready repos?

Would love to hear what’s worked (or failed) for others.

https://redd.it/1krwa4y
@r_devops

What happened to DevOps Paradox podcast?

No new episodes for ~3 months, any ideas about what happened to Darin and Victor?

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

Want to pivot into DevOps

I am a senior technical support engineer with 20 years of I.T. experience. I have been around the block, road hard and put away wet... I want to pivot into DevOps as this seems to be where my career path is taking me. My skillset is strong with Networking, Linux, Docker, Azure, any Cisco crap along with Palo Alto crap, some programming like SQL and very little python and just super strong troubleshooting skills just from being in the field for so long. I really hate certifications but I do have AZ900 and Sec+ but I do not think they matter for me with my experience and also degree.

I am a very good interviewer and can sell myself well and answer any technical question thrown at me. My question is what skills should I learn and master to add to my skilltree? More Python? Do I have to start at the bottom with junior DevOps roles? I should be able to look into more senior roles with my experience in IT?

https://redd.it/1ks28g8
@r_devops

Are we heading toward a new era for incidents?

Microsoft and Google report that 30% of their codebase is written by AI. When YC said that their last cohort of startups had [95% of their codebases generated by AI](https://leaddev.com/hiring/95-ai-written-code-unpacking-the-y-combinator-ceos-developer-jobs-bombshell). While many here are sceptical of this vibe-coding trend, it's the future of programming. But little is discussed about what it means for operation folks supporting this code.

Here is my theory:

* Developers can write more code, faster. Statistically, this means more production incidents.
* Batch size increase, making the troubleshooting harder
* Developers become helpless during an incident because they don’t know their codebase well
* The number of domain experts is shrinking, developers become generalists who spend their time reviewing LLM suggestions
* SRE team sizes are shrinking, due to AI: do more with less

Do you see this scenario playing out? How do you think SRE teams should prepare for this future?

Wrote about the topic in an article for LeadDev [https://leaddev.com/software-quality/ai-assisted-coding-incident-magnet](https://leaddev.com/software-quality/ai-assisted-coding-incident-magnet) – very curious to hear from y'all on the topic.


https://redd.it/1kry990
@r_devops

Cannot get GitHub Actions build to work with protoc

I've got a Rust build that needs access to protoc (the Protobuf compiler). I set it up like this:

  build-test-deploy:
    runs-on: ubuntu-latest

...

      - name: Install protoc
        run: sudo apt-get update && sudo apt-get install -y protobuf-compiler

      - name: Test
        run: |
          which protoc
          export PROTOC=/usr/bin/protoc

In addition, env has

env:
  AWS_REGION: "us-east-2"
  ...
  PROTOC: "/usr/bin/protoc"

'which protoc' outputs as expected: /usr/bin/protoc

Yet the build fails with this:

  Error: Custom { kind: NotFound, error: "Could not find `protoc`. If `protoc` is installed, try setting the `PROTOC` environment variable to the path of the `protoc` binary. To install it on Debian, run `apt-get install protobuf-compiler`. It is also available at https://github.com/protocolbuffers/protobuf/releases  For more information: https://docs.rs/prost-build/#sourcing-protoc" }

I'm kind of at a loss...

https://redd.it/1ks3i2a
@r_devops

Can Gitlab’s native ‘Dependency Proxy for packages’ feature replace the need for Sonatype Nexus?

Based on a developer's feedback, there's a clear need for an internal binary repository within our network to serve as a secure, controlled intermediary for external dependencies. We currently have the following issues:

1. Manual downloading, scanning, and internal placement of dependencies is time-consuming.

2. Current development workflows are being hindered by lack of streamlined access to dependencies.

3. We have no way to externally source NPM packages and NuGet packages into our environment without going through a tedious manual process.

I was looking at Gitlab’s documentation for the Dependency Proxy feature but there is no clear example of a user proxying the flavor of packages I am interested in the way you would during a build if you had Nexus or JFrog. YouTube videos around this feature are YEARS old by the way with no examples for doing this. I think we need Nexus so we can scan the proxied packages for vulnerabilities, but I would like to save cost using any workarounds in Gitlab (what we have) if that is possible.

This is apart of an ongoing effort to modernize multiple applications (running them as containers in a VKS cluster), but it doesn’t make sense to move on to this step if we have no central space for storing container images (I am aware each project in Gitlab can store container images at the project level), binaries, externally sourced dependencies that are scanned and other artifacts.

https://redd.it/1ks8718
@r_devops

What do you wish someone told you when you became a DevOps engineer?

Hello all,

What do you wish you knew when you got started in DevOps?

A tool you saw someone use every day that you adopted, a monitoring platform you switched too later than you should have in hindsight, a solution to a problem you didn't know you had, etc.

I recently got promoted internally from Systems Administrator to DevOps(yay!). I have a background in Linux/cloud administration.

I've basically been doing both systems administration and DevOps for a couple years for my company. Which means I haven't been able to do either as well as I would like.

We're bringing on a SysAdmin this week and I was moved to DevOps. So now I will have the space to do this job properly.

our stack is:
AWS:
\-ecs(fargate)
\-s3
\-guardduty
\-eventbridge
\-sns
\-route53
\-cognito
\-ecr
\-cloudwatch
\-IAM


DB:
\-mongodb atlas


Monitoring:
\-newrelic


Some things I have already identified:
I already know we need to lower our attack surface, I think we're leaving some things on the table with GH's automation(we already use GH but there's more stuff we could do with automatic tagging for issue tracking), Im planning on creating a web portal so my developers can turn on/off dev tenants as needed(ecs fargate + terraform + authenticated web portal via cognito with org SSO), and im planning on ramping up our underutilized new relic implementation and cloudwatch.

https://redd.it/1ks8f35
@r_devops

How do you avoid CI and CD unsync when using GitOps workflow like FluxCD?

Imagine situation: you push changes into the GitLab repo, docker build+push runs for 5 minutes. The FluxCD checks the repo for changes every 1 minute.

You merge a feature into the main, starting the CI/CD workflow of deploying to the production K8S. But the problem is that FluxCD is simply checking every 1 minute the repo for changes, and it triggers its deploy faster than the docker image building stage in the registry.

Is there a way to configure FluxCD to avoid such race condition of mismatched image build and deploy timings? Or should I make the FluxCD deploy only specific image hash, and bumping it to the new image manually?

https://redd.it/1ks9ola
@r_devops

FREE GitHub Advanced Security Certification

Just wanted to share a great free opportunity from GitHub for anyone

How it works:

Step 1: Complete 3 GitHub Skills courses (each ~1 hour)

Step 2: Submit the Completion Form
After finishing all three, fill out the official form to share your progress.
Deadline: May 31, 2025

Step 3: Take the Certification Exam
In June 2025, you'll receive a free voucher (worth $99) to take the GitHub Advanced Security Certification exam. If you pass, you'll earn an official GitHub certification to showcase your security skills!

I think this is a solid opportunity for anyone looking to boost their cybersecurity portfolio especially if you're interested in DevSecOps

Link:
https://maintainermonth.github.com/security-challenge

Don't forget to upvote :)

https://redd.it/1kscd08
@r_devops

Suggested resources for starting as a junior devops engineer

I’m starting as a junior devops engineer soon and was wondering if some people could point me to resources to help me get started. For background, I am currently a software engineer but in the robotics/automation field so the job I’m switching too is a role that will be relatively unfamiliar to me. I am good with Linux and python but haven’t used AWS systems or kubernetes which are what I will be working with. There will be on the job training but I don’t want to go in totally blind.

https://redd.it/1ksd89e
@r_devops

Where to apply for Internships and Jobs ?

Hey, I am a student in my final year exploring and learned DevOps , cloud, IaC, Development. I am currently applying for internships on internshala portal but I lack some skills mentioned in the requirements which I am working on right now .
I just wonder if anyone could recommend some best portals or sites to apply.

https://redd.it/1ksdoo9
@r_devops

Am I capable of junior DevOps Engineer roel with this experience ??

morphing personal info for safty

Experience:
Devops, Intern, company.ai - company networks Project January 2025 – present

• Implemented SigNoz for Kubernetes cluster monitoring, configured 30+ alerting mechanisms, and designed 5 types of dashboards for comprehensive metric visualization.

• Integrated Trivy (DevSecOps tool) with GitHub Actions, enabling automated security scans and identifying 15 high-severity vulnerabilities before deployment.

• Troubleshot Kubernetes clusters, leveraging ArgoCD and Helm charts with Horizontal Pod Autoscaling (HPA), resulting in a 25% improvement in deployment stability and optimized CI/CD pipeline efficiency

\---

Software Engineer, Intern, company2 Project June 2024 – August 2024

• Integrated NFT APIs with the frontend for dynamic asset displays, optimizing data retrieval, reducing redundant API calls by 70%, and improving API response times from 2-3s to 350ms.

• Configured Moralis and Infura for secure NFT transactions and blockchain interactions, achieving a 95% transaction success rate and reducing gas fees by 20% through smart contract execution (average execution time reduced from 4s to 2.5s)

\---
Skills :

Java, Python, NodeJS, HTML5, CSS3, Linux, SQL, Docker, Kubernetes, Git, CI/CD, Azure Cloud, AWS, Grafana, Prometheus, Signoz

\---

Projects

1. Fusion Linux - Linux Distribution for DevOps And Cloud Environments

• Automated ISO image creation and customization using live-build, Bash scripting and other configurations

• Implemented CI/CD pipelines (GitHub Actions/GitLab CI) for automated OS builds and testing, decreasing deployment time from 45 minutes to 20 minutes and improving build success rate to 98%..

• Enabled GPU passthrough for virtualized environments, improving computational performance by 90% for GPU-intensive workloads in virtual machines.

2. Infrastructure Monitoring and Vulnerability Scanning Suite | Signoz

• Monitoring solution using Signoz

• Configured 30+ custom alerting rules and developed 5 types of dashboards, improving system observability and reducing mean time to detect (MTTD) by 40%.

• Integrated Trivy for automated vulnerability scanning in containers and system packages, identifying 15+ high-severity vulnerabilities per scan and reducing security risks by 60%.

3. Cryptway | React Js, Rapid Api, Solidity, Ethereum, Vercel

• Developed a blockchain platform enabling users to create Ethereum wallets, send/receive Ethereum, and swap ERC-20 tokens, processing an average of 080+ transactions per day

• Migrated from Vercel to Azure Cloud for enhanced scalability and cost optimization, leveraging Azure Spot Instances to reduce infrastructure costs by 70% while maintaining performance.

\---

Achievements

• 1st Prize at Mumbai Hacks Hackathon (World’s Largest Generative AI Hackathon)

• Smart India Hackathon Finalist 2024

• 1st Prize at AI Spark (Hackathon)

\---

Certifications

• Microsoft Certified: Azure Fundamentals

• Microsoft Certified: Azure AI Fundamentals

az 104

and preparing for CKA

https://redd.it/1ksk7li
@r_devops

AWS project

I would like to make an AWS project that would basically help me explore what I like and what I don’t like. I’m pretty new to public clouds but I’ve got experience with onprem so the learning curve is not that steep. I was suggested to do something like an app to call taxis. Does anyone have any other project suggestions that would force me to not only write code, but also do infra, security and data management related things?

https://redd.it/1ksl38t
@r_devops

What's your favorite lightweight monitoring stack?

Prometheus feels a bit heavy for small projects. Any go-to minimal setups you like?

https://redd.it/1ksm95r
@r_devops