Making Sense of Cloud Spend

Hey y'all.. Wrote an article on sharing some throughts on Cloud Spend

https://medium.com/@mfundo/diagnosing-the-cloud-cost-mess-fe8e38c62bd3


https://redd.it/1k7eabl
@r_devops

Devops workflow tips for a frontend application developer who needs to take on more ops responsibilities.

What is an efficient workflow/work environment setup to tackle an ops task that involves a Github 'Action', and a Bitrise build 'Workflow'.

I've written the GitHub Action as a bash script, and the Bitrise Workflow is a collection of pluggable Bitrise 'Steps' and some custom scripts in the repository that are triggered from the Bitrise Workflow.
The GitHub Action responds to the creation of a new tag with a name that matches, and the Bitrise Workflow runs build tasks that call our backend REST API for dynamic configuration specifics.

I find working on the ops stuff outside the monorepo slow and inefficient.

* Re-running scripts on remote machines/services is slower (I run the service using their local client to debug, but it's difficult to replicate the VM environment accurately in my local machine)
* They often break because I miss mistakes in the bash scripts (don't have editor/language based tools to help me here)
* The cloud based builds need time to execute because the VMs need to setup everything every time (I've cached some stuff but not all)

**Can I please get some tips on how to work more efficiently when working on processes that are distributed across systems?**

For context, I'm usually a frontend app developer and I've set up our monorepo to make our lives as easy as possible:

* Typed language (TS) and linter so we can see our errors in the editor as we work
* automated unit test runner with a 'watcher' that runs on 'save' to make sure our application logic doesn't get broken
* integrated testing pipeline that runs upon creation of pull requests
* hot module reloading so that we can visually see the results of our latests changes
* separation of presentational components and application logic with strict architectural guidelines to keep things modular
* monorepo tooling with task-runner to enable the above

**What are some devops techniques to achieve the same type of workflow efficiencies when configuring processes that run across distributed systems?**

I suspect that I need to look into:

* Modularizing logic into independent scripts
* Containers?

Anything else?



https://redd.it/1k7he32
@r_devops

ServerlessDays Belfast 2025 – “Serverless is Serving” (Thursday 15th May)

Hey folks 👋

We’re excited to announce that ServerlessDays Belfast is back for 2025! Mark your calendars for Thursday 15th May, and get ready for a full day of talks, learning, and networking—all centered around building confidently and excellently with serverless technologies.

📍 Venue: The stunning Drawing Offices at Titanic Hotel Belfast
🎯 Theme: Serverless is Serving – building with confidence and excellence
🎟 Tickets: £60 (includes breakfast, lunch, and snacks!)
Group discounts available!

This year’s focus is all about how serverless empowers developers, teams, and communities by removing the ops overhead and letting us focus on delivering real value. Whether you're a seasoned cloud engineer or just curious about getting started with serverless, this event is for you.

Expect talks from local and international speakers, including Patrick Debois, the father/grandfather of DevOps! Expect real-world stories, innovative builds, and practical techniques that show how far we’ve come since the early days of serverless. It’s not just about infra anymore—it’s about service.

🙌 A massive shoutout to our sponsors for making this possible: AWS, EverQuote, and G-P
👥 Proudly organised by volunteers from AWS, G-P, Kainos, Liberty IT, Workrise, Rapid7, EverQuote, and The Serverless Edge.

Come for the talks, stay for the community.

💻 More info & tickets: https://serverlessdaysbelfast.com/
Got questions? Drop them below.

Hope to see you there!

https://redd.it/1k7h8ug
@r_devops

Setting up DevOps pipelines is my worst nightmare

Sorry for the rant, but I need to let off some steam. I’ve been building and running cloud stacks for some years now, and it still amazes me how terrible the whole process is—no matter the provider.

You’ve got your application, you start fresh with a new template and a new cloud account (clients finally wants to migrate to the cloud). You set up your CI/CD pipeline, and the goal is to have it provision your resources in the end. You write your first draft, push it, wait for builds/tests/linting/etc... and then it hits the final step: deployment. And italways fails.

Something's broken. You missed a dependency. The runner or the deployment principal doesn’t have the right set of permissions. No one can tell you exactly what permissions your final principal needs. So you enter this endless loop of trial and error. You could skip some of that by just granting full admin rights—but who wants to do that?

Resources get created, the deployment fails but fails to clean up properly. You need to manually delete things. But wait—some resources depend on others, so you can’t delete X before Y is gone. Meanwhile, your stack is a half-broken mess, and you're deep in a cloud console trying to figure out which dangling part is blocking the cleanup.

Hours gone. Again.

You feel like you’re so close every time—just one last permission tweak, one last missing variable... but wait, are those variables even passed correctly from the CI template to the container to the deployment script?

Error messages? Super cryptic. “Something failed while deploying your stack.” Thanks. “mysql password requirements not met.” Wait—there are password requirements? Where’s that documented? Oh, it’s not in the main docs. It’s in one of the five different documentation sets—SDKs, CLI tools, Terraform providers, custom template languages... each with just enough difference to make you scream.

And the worst part? I love cloud-native development. I’m a big fan of serverless, and I genuinely believe in infrastructure-as-code. Once it’s up and running, it’s amazing. But getting there? It still feels outdated, clunky, and overly complex. It’s the opposite of intuitive.

I’m used to fast (almost instant) feedback loops when developing applications on my local machine. AI tools give me huge productivity boost. But CI/CD? It’s still “make a change, wait minutes (or hours), get an error, repeat.” It kills motivation.

And don’t even get me started on the environmental cost of spinning up and tearing down all these failed resources, countless hours of pipeline runs that fail on the last step - deploy...

Anyway, rant over. Just had to vent because this cycle has been getting to me. Same problems across AWS, Azure, GCP. Anyone else feeling this pain? Got any strategies to make it suck less?

https://redd.it/1k7j1bv
@r_devops

Off shore DevOps engineers - possible?

I started working in tech support for an American company, but I work from Europe remotely.

The possibilities of getting into DevOps from support are real, there are several people who started in support and then moved to DevOps / SRE / QA.

During interviews, the hiring manager told me that in 1.5 to 2 years it is expected for me to become either a senior supporter or move into engineering.

What wasn’t clear to me at that time, is that only customer care teams (including mine) are spread in different time zones. All engineers are in USA or similar time zones (south/central America), including DevOps team.



I am trying to understand if progression into engineering is actually possible or I have been lied to.

In support we have a global “standup” that is at 17 for the EU and 8/9 for the American supporters (they are all in east coast).

But, the devOps team is made of 3/4 folks, al in USA. They seem to have a short stand up at my 18:30 and then other meetings later.



So the question for you: say you are in a 4 people DevOps team. Would you consider a person working mostly asynchronously in your team? I wanted to trust my hiring manager at first, but my hopes went low when I realized that all the engineers are in USA time. I find it hard to believe that DevOps engineers would be up to change their times for me, and that things could work given also the need to sync and meet with developers, SREs etc in real time.



The support team uses a “follow the sun” strategy while devOps (and all engineering teams) doesn’t seem to use, want or need different hours coverage. I would also be a junior devops if I was to move into their team, so possibly in need of mentoring.



I also am very new and already asking my manager questions about leaving the team doesn’t feel right.



So.. am I cooked and have I been lied to by the hiring manager so that they could fill the position asap?

https://redd.it/1k7ioaa
@r_devops

Anyone looking for a part-time devops/consultant with previous startup experience?

Hi,

If there’s anyone US or UK based that’s looking for a part time devops, I would be open for discussion.

I was a part of several startups, one of which skyrocketed and got acquired, and the other made it to some big fancy investments. We can talk more if anyone is interested in having me on their team.

I’m open to both engineering and consulting.

Best regards!

https://redd.it/1k7km7v
@r_devops

Looking for DevOps feedback

Hey all, I'm a developer @ Korbit AI and I was hoping to get some feedback from QA / Dev Ops engineers as to how we can make our reviews even more useful for this specific type of focus.


Currently we focus on these 8 categories: Functionality, Security, Performance, Error Handling, Readability, Logging, Design and Documentation.


My question is, as a dev ops engineer / qa, what are specific types of things our reviews can really focus on to help save time in this particular subject. We're planning on releasing a new feature called Korbit Policies, where you are able to tell Korbit specific things to flag ( example is like refactoring from one class to another and enforcing usage ).


Let me know and thank you in advanced.

https://redd.it/1k7r4n1
@r_devops

API Sprawl - issue for you or na?

Do y'alls bosses see API sprawl as a real problem? Or is just your problem? We need more discoverability for our APIs for sure, too many people doing too many things off in the corner. But I also need to make sure my boss sees it as a legit issue so that I can do something about it.

https://redd.it/1k7w38d
@r_devops

Hiring: Cold Email Deliverability & DevOps Specialist (High-Scale Infrastructure)

We're looking for a DevOps/Deliverability expert who lives and breathes cold email infrastructure.
Not just someone who's familiar — but someone who's built high-deliverability SMTP servers, optimized inbox placement at scale, and knows how to get emails delivered no matter what.

This is a right-hand role — not just a task-based position.
You'll be working directly with the founder to build, scale, and optimize our email infrastructure.

Who we're looking for:
✅ Deep experience managing cold email SMTP infrastructure (PowerMTA, Postal, Mailcow, etc.)
✅ Proven ability to hit the inbox at scale — across 100+ VMs and IPs
✅ Strong DevOps/sysadmin background — building scalable, redundant systems
✅ Hands-on experience managing IP reputation, rDNS, DKIM, SPF, DMARC, smart routing, etc.
✅ Creative problem solver — can build and adapt systems to changing deliverability challenges
✅ BONUS: You've built email warm-up and/or email verification tools (especially catch-all detection)

What you'll be responsible for:

Architecting and managing a large-scale cold email infrastructure
Developing an internal warm-up tool (we'll provide resources to accelerate it)
Building an internal email verification tool (with catch-all logic, bounce detection, etc.)
Managing 100+ VMs/IPs and continuously improving deliverability rates
Innovating ways to stay ahead of spam filters, blacklists, and reputation risks
Making sure we scale without deliverability or infrastructure breakdowns

What we are:
We're an email infrastructure software company built specifically for cold email at scale.
We care deeply about quality infrastructure — and we need someone who gets it, fast, and can build it right.

What we’re NOT looking for:
❌ Someone who's only run basic transactional email servers
❌ Someone who needs step-by-step instructions
❌ Someone without experience running cold email systems at serious scale

If you're a builder who loves solving tough deliverability problems and wants to create something massive — let's talk.

https://redd.it/1k7wok3
@r_devops

What are your pain points in debugging kubernetes deployments?

The biggest pain point I have seen a lot are those frustrating scenarios where "everything looks healthy" but your system isn't working (like services not talking to each other properly or data not flowing correctly).

Would love to hear your debugging pain points and how we could make this more useful. Is this something you'd find valuable?

https://redd.it/1k7y9zi
@r_devops

Learn how to debug SQS consumers in Kubernetes without rebuilds

Debugging SQS consumers in Kubernetes isn't for the faint of heart. This guide shows how you can debug them locally using mirrord queue-splitting model, without disrupting production consumers.

Hope it will help you save some precious time =)

https://metalbear.co/guides/how-to-debug-sqs-consumers/?utm\_source=organic\_social&utm\_medium=reddit\_organic&utm\_campaign=reddit\_post

https://redd.it/1k7zrx5
@r_devops

Created DevOps Project... real-world, hands-on, esp. useful for people who look for a job.

I created hands on DevOps project to help people looking for a job or upskill to fill the gaps in practical knowledge.

I recently did bunch of interviews and I think it will help a lot. Even if you don't have time to do it, just go through the content, it is free. Now I know there are some things that are not covered there, but still it is great foundation for about 70% of daily tasks.

It is close to what is used in most of the companies I worked (but trimmed down to save resources). It is fully hands on, you build app, containerise, deploy, create ci/cd, template with helm, use kubernetes, use terraform and aws, create monitoring and list goes on..

here is the video where I talk about it: https://youtu.be/vtCW5IgJ9-A?si=8nfBu4vgN4uhdX-2

here is the project itself: https://prepare.sh/project/devops-foundational-project

https://redd.it/1k80zlj
@r_devops

Tool for docs generate and host

What tool you use for publish documentation ?like do docker kuber and etc

Now I have cicd what copy readme.md in one central project docs with versions by tag .

https://redd.it/1k81lmy
@r_devops

Blind posts are crazy

Guys, have you checked recently the Blind posts about job offers? Just went through some of the very recent posts and felt like we live in different dimensions. When here I see a lot of people struggling even to land an interview for a long time, some even for 2 years despite being experienced those guys are on the fence between, or even among a gargantuan TC offers. One guy posted about having 3 offers (Databricks, Meta, Google) on the table, with tremendous TC, and was looking for some second opinions, etc. It’s really crazy.
Of course, I’m happy for every single person who gets an offer, but at the same time, I feel sad for others who are struggling.
What is this gap about? There is no balance. Why do we have such a huge abyss between the communities in the same geolocation? What do you think about it?

https://redd.it/1k84mq9
@r_devops

What do you tell non technical people what your job is?

Title says it all.

https://redd.it/1k87uwj
@r_devops

From mobile dev to devops

Hello, I’m new here. Lately, I’ve been browsing Reddit to understand how hard the transition from software developer to DevOps is. I noticed that most people making the switch come from a backend background. I’m a native mobile developer with 2 years of experience, and I’m wondering—how difficult would it be for someone like me to move into DevOps? Would my experience be considered valuable, especially if I build DevOps projects on the side? Would HR see me as a good fit? I’d love to hear your thoughts.


https://redd.it/1k87xj4
@r_devops

Kubetail: Real-time Kubernetes logging dashboard, now with Search

Kubetail is an open-source, general-purpose logging dashboard for Kubernetes, optimized for tailing logs across multi-container workloads in real-time. The primary entry point for Kubetail is the kubetail CLI tool, which can launch a local web dashboard on your desktop or stream raw logs directly to your terminal.

I started working on this project two years ago after getting frustrated with the Kubernetes Dashboard's log viewer and I'm excited to share that we’ve added some new features, including search!

# What's new

# 🔍 Search

Now you can grep/search your container logs in real-time, right from the Kubetail web dashboard. Under-the-hood, search uses a super fast Rust executable that scans your raw log files on-disk in your cluster, then sends only relevant results back to your browser. Now you don’t have to download all your log records just to grep them locally anymore. The feature is live in our latest release candidate - try it out now here: https://www.kubetail.com/demo.

# 🖥️/🌐 Run on Desktop or in Cluster

Kubetail can run locally or inside your cluster. For local use, we built a simple CLI that starts the dashboard on your desktop (quick-start):

# Install
$ brew install kubetail

# Run
$ kubetail serve

It uses your local kubeconfig file to connect to your clusters and you can easily switch between them. You can also install Kubetail inside a cluster itself and access it from a web browser using kubectl proxy or kubectl port-forward (quick-start).

# 💻 Tail logs in the terminal

Sometimes you can't beat tailing logs in the terminal, so we added a powerful logs sub-command to the kubetail CLI tool that you can use to follow container logs or even fetch all the records in a given time window to analyze them in more detail locally (quick-start):

# Follow example
$ kubetail logs deployments/web --follow

# Fetch example
$ kubetail logs deployments/web \
--since 2025-04-20T00:00:00Z \
--until 2025-04-21T00:00:00Z \
--all > logs.txt

# 📐 Clean UI

We’ve worked hard to make Kubetail feel fast and intuitive. One feature that our users love is that multi-container logs are merged into a single timeline, color-coded by container—so you can track what’s happening across pods at a glance. Using simple controls you can quickly go to the beginning of the merged timeline, tail the ending, or scroll through the event timeline. Our goal is to make the most user-friendly Kubernetes logging tool so if you’re passionate about design and you love logs, we’d love your help! (Thanks victorchrollo14 and HarshDeep61034 for your recent contributions!)

# 🎯 Easy filtering

When something’s on fire in your cluster, you need to quickly isolate the issue—whether it’s tied to a specific region, node, or pod – so we added quick filters to help you narrow the log sources you're looking at. You can also filter by time to quickly narrow your debugging window to around the time an incident occurred. Soon we're planning on adding more filtering options like labels too so you can create your own groups of pods to filter on.

# ⏱️ Real-time

One of my original frustrations with the Kubernetes Dashboard is that it refreshes container logs every few seconds instead of just streaming data as it comes in, so we built Kubetail to be able to handle data in real-time. In the Kubetail web dashboard you can see messages as soon as they get written to your cluster. Kubetail also subscribes to messages from new containers automatically as soon as the container is started so you can track requests seamlessly as they jump between ephemeral containers even across workloads. That means I don’t need to keep multiple Kubernetes Dashboard

logging windows open any more!

# 🌙 Dark Mode

We didn't want users to get blinded when they opened up Kubetail, so we added a dark mode theme that picks up on your system preferences automatically. Hopefully streaming logs lines will be easier on the eyes now.

\---

If Kubetail has been useful to you, take a moment to add a star on Github and leave a comment. Your feedback will help others discover it and help us improve the project!

\---

Join our community on Discord for real-time support or just to say hi!

https://redd.it/1k8arks
@r_devops

How difficult is the process for publishing an app to the Android and Apple Store?

Hello All,

I've been working on a mobile game and am going to release it to the app store at some point.

I had a couple of questions about app publishing.

1. How much time does app publishing process take? Is it a lot of work? Seeing compliance lists such as https://developer.android.com/docs/quality-guidelines/core-app-quality#sc intimidates me.

Are they actually enforcing all these rules?

2. I see there are tools available like Runway, Tramline, FastLane that claim to make the deployment and publishing process easy.

Have any of you used these tools?

Do they help reduce time to publish and update or would I be better off writing scripts/github actions for this?

3. ⁠Do you know any tools that automate all this compliance stuff away?

Thanks a lot :)



https://redd.it/1k8bft1
@r_devops

The Easiest Way to Manage Multi-Container Apps (Perfect for Small Projects!)

Hey everyone! As part of my 60-Day ReadList Series #4: Simplifying Docker & Kubernetes.

This time, I break down Docker Compose. How it simplifies managing multi-container applications, Why it’s so useful, How to structure a docker-compose.yml, and some bonus tips like scaling, using environment variables, and networks.

Covered topics include:
1. Why Docker Compose is a must-have tool
2. Breakdown of docker-compose.yml structure
3. How volumes help persist container data
4. Scaling services with a single command
5. Managing environment-specific configs
6. Networking between containers

Perfect for someone who’s starting out with Docker and building small projects. Docker Compose handles things surprisingly well without the heavy lifting!

If you’ve been wanting to get more comfortable with Docker and want a beginner-friendly guide that’s actually practical, check it out. Docker Compose Made Simple: Deploying Multi-Container Applications in Minutes

Thanks for reading and supporting the series!

https://redd.it/1k8bdzu
@r_devops

Is it normal to feel overwhelmed at a new DevOps Job?

Hello, I just joined a multinational company. Their infra has already been setup and has fully matured. I feel overwhelmed on the stuff I have to learn and teams to communicate requests to, not to mention transitioning from unix terminals (Used to live in the terminal) to windows (Restrictions).




Some info about me, previously worked from a startup and previously a mid sized company (That also came from a startup). It was easy learning and building the infra of the two. And right now, I feel so weak.




Lemme know if you guys have any advice, I would highly appreciate it.



https://redd.it/1k8di8q
@r_devops