What tools should I use to setup a portable CI/CD pipeline for four repositories towards a single VPS?

Hey folks, I'm currently part of an open source club at my college and recently I convinced our members to move and streamline all of our infrastructure to a single VPS (it's currently hosted all over the place across different IAAS's and PAAS's and is a pain in the ass to manage and troubleshoot)

While I do run a homelab and have some minor experience with GitHub actions and docker, I'm pretty new to devops and automating stuff. My current approach with one of my personal projects is to run a GitHub action which SSH's into my homelab and runs a few commands to pull the changes and restart the service.

The nginx configs, envs and systemd services were all written by hand and while I could do this for the club projects, it feels quite inefficient and ideally I would like a system which we can quickly migrate to a different VPS if the need arises.

I've done some research but I'm quite confused with which combination of tools is best suited for this job, initially my plan was to dockerise all the projects and create a GH action which builds the containers and pushes it to a registry. I would run watchtower on the VPS (which i would setup using Ansible) to detect these pushes and automatically update the containers but then I saw in the watchtower readme that it is not meant to be used in production environments, and they instead suggest to use Kubernetes which is something I'm not sure I require since I only have a single server.

While I know my needs aren't quite "production" worthy, I would like to use the right tool for the job and learn something useful in the process. Does it make sense to learn Kubernetes for this and does my initial approach make any sense or do I need to go and do some more research? Any pointers are appreciated!

https://redd.it/1g0fhr3
@r_devops

Getting into a Devops as a FullStack developer

As the title says, I have been working as a Full-stack web/Mobile developer for 3+ years, I want to try something new, so I choose Devops for my next transition. I worked primarily on Golang & Node.js. There are a lot of tools available for DevOps but still, I'm not sure how to start from. I don't like to watch videos for countless hours for learning, so kindly recommend project-based courses, blogs, or any articles that would help. I appreciate your valuable comment.

https://redd.it/1g0ez3r
@r_devops

30 Days Of CNCF Projects | Day 3: What is KEDA + Demo ↔️

Hey all! 👋

I’ve just published a new video for my project 30 Days of CNCF Projects, and this one’s all about KEDA – Kubernetes Event-Driven Autoscaler! In this video series, I’m diving each week for another CNCF project.

In the video, I cover:

1. What KEDA is
2. How it works
3. A live demo + a workshop to try it yourself

Check it out here: https://www.youtube.com/watch?v=VlrEXJdEc5w

This is part of my 30 Days of CNCF Projects series, where I explore different CNCF tools to help us solve real-world cloud-native challenges. Would love to get your feedback on the video and hear if anyone here has worked with KEDA! 🚀

Looking forward to your thoughts! 😃

If you want to help me promote this promote, I will be grateful for it.

Like, share, subscribe & connect me on Linkedin - https://www.linkedin.com/in/guy-menahem/

https://redd.it/1g0f5an
@r_devops

todegree or not todegree?

Hello everyone, I'm 19 yo, I've been studying CS for quite a while now, Built some fullstack projects and have a decent understanding of the basics and still learning more and still have a lot to learn.

Now I'm stuck between going to collage or not,

If I started going to collage
1 - I would have to work full time, to pay fees.
2 - More work mean less study time, I used to work full time for a long period of time and managing work study time is always hard, like it would be great if studied 1 hour a day.
3 - I'm living in 3rd world country and the education system is dog shit I will just go there for the degree since its mostly required and I will have harder time finding job without it.

If I didn't
1 - currently I'm studying 4-6hrs a day, the internet is full of CS studying materials, And so far I'm doing great.
2 - I could learn faster therefor build more projects and this could get me a job and prevent time loss.
3 - I may have harder time finding a job science it mostly require a degree

I hope someone with more experience help me through this.

https://redd.it/1g0mdmr
@r_devops

Do you store secrets in environment variables?

Surely, all the tutorials and user docs across tools use code examples like `process.env.OPENAI_TOKEN` and other such examples. So yeah, it is pretty common and it also easily spills to developer projects.

How do you manage these secrets in your team projects? how do you balance a solution to the problem that is both secure but also provides nice DX to developers and doesn't antagonize them?

I wrote a very lengthy blog post about all the reasons I could think of to COMPLETELY AVOID secrets in env vars and my proposed approach. Happy to learn what you all are doing in practice and how to improve on my go-to best practices.

https://redd.it/1g0muvv
@r_devops

Restarting My DevOps Career in the US: How to Maximize 3 Months?

Hi everyone, I moved to the US on a marriage visa just under a year ago. Back in my home country, I worked as a Linux admin for 1.5 years, then transferred to a DevOps role for 2 years before coming to the US. I can honestly say that my first year here has been quite challenging, especially due to the language barrier and being neurodivergent.

For the first 6 months, I focused on preparing my resume, applying to over 200 jobs, and prepping for both behavioral and technical questions. I interviewed with three companies. I was brave enough to jump into it, but the interviews made me realize that while I have impressive project experience, 1) a lot of time has passed, and my memory of the details has faded, 2) I lack basic IT knowledge, and 3) my English has become wordy, and I struggle to find the right expressions. These experiences left me with interview anxiety, which caused something like panic attacks. I’ve been focusing on personal recovery but am now ready to restart my job search.

Goal: Get hired during the hiring season (January/February).

Current Status:

1. Continuously improving my resume.
2. Relearning basic IT knowledge (KodeKloud - DevOps Engineer Learning Path, mastermnd - DevOps Bootcamp).
3. Writing down interview answers.

Strengths (as pointed out by former colleagues):

1. Excellent at troubleshooting—always manage to make things work.
2. Great communicator and mediator.
3. English used to be my strength... but I guess not anymore.

Weaknesses:

1. No experience with Infrastructure as Code (my former company had a separate team for that).
2. Only experienced with GCP.
3. No CS degree—my background is mainly in Ops, which I worked in for just one year.

I now have roughly 3 months left to achieve my goal. If you were in my shoes, how would you spend these 3 months? Any advice would be greatly appreciated!

https://redd.it/1g0p896
@r_devops

From 0 to 3k PODs in "notime"

interesting what is possible when you know what you are doing...
https://www.youtube.com/watch?v=nQlV7hEU8ro

https://redd.it/1g0p5dc
@r_devops

CIDR is kind of kicking my @$$

Hi, all, I'm very new the world of IT and taking a networking class this semester at school. The whole CIDR thing is very confusing to me. I understand the basics of binary and getting the ranges and whatnot down, but I had a lab thrown at me where we're asked if host addresses for a network are valid or not and explain why.

I will say the professor I have is not very good at breaking things down for newcomers, he operates at a much higher level than an intro course should be, IMO. So trying to keep up with him is also challenging.

The lab gives a network ID of 192.168.5.0/24. Now, my understanding is that the /24 means the first 24 bits of the address are "masked" or locked in place, and turned "on." These are now immutable, and host addresses will need to match those first three octets to be even be up for consideration as "valid."

So the first one, for example, 192.168.6.10/24 is not valid, because of the .6 portion of the address will not match the required network address portion dictated by the CIDR notation, However, the next one, 192.168.5.10/24 is valid, since the octets match the CIDR notation, 192.168.5.x.

I really just need a solid breakdown on how to differentiate and learn this CIDR stuff, I can tell its important moving forward to understand further concepts, the prof is just not putting it together in ways that click for me.

https://redd.it/1g0ow3q
@r_devops

How do you keep ci-cd configuration file up to date between project branches?

Hi, I have multiple environments associated with their branches. Deployments are made with Deployer via one pipeline job which detect corresponding environment in gitlab-ci file. How do you keep changes made on this file and Deployer recipe up to date on each branch?
Currently I cherry-pick changes on each of the branches.
I think to create ci-cd-config branch for everything related to ci-cd and merge changes on other branches.

https://redd.it/1g0tpug
@r_devops

In the world on automation, which language gets requested the most between Python and Go? Do you think that will change?

Nothing to add to the title. It’s pretty straight forward.

https://redd.it/1g0tph0
@r_devops

The dillema: QA'd weekly releases or release ever hour with extensive integration tests?

I've run two startups for about 15 years, then joined 3 other startups that I wasn't a founder of.

When I ran my own startups I migrated to a system of extensive testing, and continuous releases.

The tests sort of made it impossible to push a broken build. When one snuck through, I'd do a post mortem analysis, and then shore up our tests to prevent that from ever happening again.

However, the startups I joined have had a terrible release process.

All of them had NO tests and just a "let's do everything very very very very carefully" approach to software engineering (which basically doesn't scale).

What ended up happening, is that once you pick this fragile deployment method, you're basically stuck.

It takes a ton of effort and changing the team to migrate BACK to continuous delivery.

The current startup I'm at has a weekly release schedule.

This costs us tons of time in lost productivity:

- more managers have to work around the timing of each release.

- engineers have to triage tickets making sure features and PRs are merged at the right time.

- it causes our PRs to be HUGE for various reasons which also snowballs into even bigger PRs because devs have to get everything into one big release.

- we don't invest in any testing so I can't personally to TDD and deliver high quality code. It makes dev a huge pain for me and it's not enjoyable.

- if we screw up, we have to wait until another push

- if an urgent bug fix happens, we have to cherrypick it and get it into prod.

Is continuous delivery accepted yet as best practice?

The issue I have now is that I'm going to try to convince our CEO that we have to make this change.

What I'm worried about here is that it's hard to migrate back to a continuous delivery practice and it will require time and effort.

If the project fails, it's going to be on my shoulders and the fingers will point towards me. Plus I'm going to have to allocate extra time and effort to make sure it works fine.

What I'd like to do is just tell him that this isn't a controversial suggestion - it's industry standard best practice.

But the question, is it? Do other startups take this approach or am I just biased because it's my personal preference?

Thanks guys!

https://redd.it/1g0vtys
@r_devops

how to manage secrets in gcp to have stateless projects?

Hi,

I have a shared project called tooling and several environment projects: dev, staging, prod

My idea is to have all dev,staging,prod stateless

But then how should I handle the secrets? we will be using mostly GKE for the apps.

I am using Terraform CDK with typescript.

My ideas are:

1. in Github actions store the secrets and in the app deployments create gke secrets replacing the map in every deployment so it does create/updates all the time and in one operation.
2. in Github actions store the secrets , create the secrets in Google secret manager at the app deployments, the problem is to handle the initial value and the updates, check if exists then create else create one version every time the app deploys? is not too many versions?
3. keep all secrets in the shared tooling project, for all the environments?

Any ideas? thanks

https://redd.it/1g0x658
@r_devops

How popular is this Wolfi base image as alternate to Alpine? Do you use it in production?

I am exploring ways to move away from Alpine as I encountered some DNS problem with it recently. Is Wolfi a good alternative base image? Please don't suggest bloated Debian and Ubuntu

https://redd.it/1g0wgky
@r_devops

Advice for new manager of a small team (3)

Hi - new manager here.

My background is SQL, python, powerBI automating data collection, creating reports and dashboards.

I used to work solo, on an island, with no real experience maintaining code for anyone but myself.

I’ve got my team a repo and we are working to deploy a pipeline and I need help with how we are going to manage our branch / branches…

Right now our plan is to spin off new branches for work items and push them into main when they’re ready.


We meet Monday + Thursday to chat formally.

I have set up coding guidelines and a repo.

What else can I do?

https://redd.it/1g0zrqo
@r_devops

Transition to Solution Architect?

I have a few years of experience in DevOps now.

I don't have a cloud cert under my belt just yet. Recently started working on getting AWS Solution Architect Associate cert.

I did a take home architecture exercise to get my current job. It was interesting and made me think, perhaps I would enjoy architecture more than DevOps. Maybe, I'm a big picture person?

DevOps is fine but I'm not sure I see myself doing this for the rest of my career.

How can I transition to a Solution Architect role? How would I know if being a Solution Architect is right for me?

Are there any Solution Architects out there that can tell me about their day-to-day?




https://redd.it/1g15168
@r_devops

A Self-Hosted Code Review and Analysis Server


We have built a self-hosted code review service, designed to be useful in the following scenarios:

* You have many repos but still want tight control over code quality
* Your repos are private, and commercial services seem overkill
* You want to continuously improve the process and rules, with full customization

We are open-sourcing it and hope it will be helpful.

[https://github.com/qiniu/reviewbot](https://github.com/qiniu/reviewbot)

Welcome feedback and suggestions. Thanks\~

https://redd.it/1g15h3s
@r_devops

After 3 Years on the Same Tech Stack, What Skills Should I Refresh for DevOps?

I've been on a consulting project with a bank for the past three years, but now that it's wrapping up, I'll be on the bench. My work has primarily involved GCP migration from on-prem using GitHub Actions for CI/CD and Terraform Enterprise for IaC and deployments. After three years of sticking with the same tech stack and mostly writing YAML, I feel like I’ve lost my edge and need to refresh my skills. Any suggestions on areas, tools, or skills I should focus on to get back up to speed?


TL;DR: Spent 3 years on GCP migration using GitHub Actions and Terraform. Project’s ending, and I feel rusty. What should I focus on to stay sharp in DevOps?

https://redd.it/1g16hkx
@r_devops

What project DevOps can build to make USA peoples say "Wow wow wee waa!"?

Jagshemash, DevOps neighbours!

It is I, Boyan, greatest DevOps in all Kazakhstan! I come to you with important question. I want to show my skill to USA companies—yes, land of McDonald’s, Pamela Anderson, and big monies! But how can I make them say, “Wow wow wee waa! This Boyan, we must hire him immediately!”?

What project can I make as DevOps engineer that is big and glorious? Something that will showcase all my big brain powers and make US and A recruiterka slide into my DM like smooth homemade rakiya.

Here is what I know to do very nice:

Make pipelines go fast, like rocket on cow’s milk.
I best snake handler in village: mostly pythons.
Automate things so I can rest and eat more cheese while servers run themselves.
Kubernetes? Yes, I can do! Even my neighbor Nursultan say, “Boyan, you are kuber-whatever genius!”
I also do monitoring, alerting, and can fix everything with only 3 lines of code—maximum!

So what can I build? Maybe I make:

Big project with CI/CD pipeline that deploy faster than gypsy stealing chicken?
Or I make kubernetes cluster that self-heal like strong Kazakh man?
Or maybe cloud infrastructure that so big and scalable, it can hold all of Kazakhstan’s goats?

What will make hot recruiter lady say, “This Boyan, we need him on remote contracts, fast!”? Please help me, friends! I want to bring my glorious DevOps talent to America!

Chenquieh!

Boyan Balgaran, soon-to-be American DevOps superstar

https://redd.it/1g173ww
@r_devops

Naming conventions for VMs?

Hello,

New to DevOps. Just started this role less than a month ago.

I am being tasked currently with writing up terraform for the existing infrastructure that was created through the cloud provider WebGUI, and with that I’m being tasked with coming up with a naming convention for these instances since there isn’t really any consistency between them. I have to account for environment, and scale.

So- I’m thinking most general -> least general, so these instances are grouped alphabetically by their env essentially.

For example- dev-app-01 or something.

Do you guys have any recommendations? Any tips or advice?

https://redd.it/1g17eco
@r_devops

Cloud and devops vs ml


Currently in 3rd sem been doing web dev for 7 months , I am not that good in web dev as of now but for long run I am thinking to do cloud and devops after web dev

don't have any prior knowledge of ml so it would be totally new as for future what should be my goal to learn after web dev should it be cloud or ml

I don't have a clear goal as to what to do I am just learning tech stacks and all and am bored doing web dev so thinking of switching to something else

https://redd.it/1g172ct
@r_devops

If you're struggling to learn, we have a bunch of projects!

Hey everyone, I'm Dan from roadmap.sh (which I know gets posted and mentioned all the time haha)!

We've been working hard on providing people with projects to help prove their knowledge, because as you know, the best way to really learn something is to build it!

We now have 21 DevOps projects that you can build with a good amount in Basic, Intermediate and Advanced!

https://roadmap.sh/projects?g=devops

If you want to see any other classic projects here then just submit an Issue on GitHub.

https://redd.it/1g19yc4
@r_devops