Unleashing the power of frame pointers pt.1

Hello everyone, I just wanted to share with you the first episode of a blog series. This series dives into frame pointers and their critical role in building a CPU profiler.

https://blog.maxgio.me/posts/unleashing-power-frame-poiners-execution-environment/

This episode will explore the fundamentals of a program's execution environment. By the end, it shows how frame pointers become essential for stack walking.

I've written it while writing a basic sampling-based kernel-assisted profiler. Indeed, the second episode will be more practical and see how to code a basic profiler, sampling stack traces and accounting them for CPU time, etc.

As we always have something to learn, any feedback is more than appreciated.

Thank you.

https://redd.it/1dn7rhw
@r_devops

Shorten your feedback loop: Java observability with OpenTelemetry, Grafana Cloud,

https://grafana.com/blog/2024/06/21/shorten-your-feedback-loop-java-observability-with-opentelemetry-grafana-cloud-and-digma.ai/

https://redd.it/1dn8m3o
@r_devops

I made a simple introductory quiz for understanding Kubernetes basics

Hey everyone, I've been learning K8s recently and I thought I'd put my basic knowledge to the test with a little quiz for myself and my team at work! I thought I'd share it with you all!

Quiz

https://redd.it/1dn9yfy
@r_devops

All the experience folks out there What's your go to strategy to switch jobs

Let me get straight to point have good skill set for 2 Y.O.E but pay is bad
Want to switch but don't know where to begin
Just before you guys say LinkedIn
The hard truth is too much crowd
I'm a devops engineer btw so all the experience folks please enlighten me on your strategy.

https://redd.it/1dn9q67
@r_devops

Can one person manage a Devops practice?

I know it’s redundant because DevOps is an alleged culture, but my real Q is, can one person admin 5+ K8s clusters. 20+ NS and build out TF templates as well as ArgoCD maintenance? How sustainable is all that?

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

The Open Source Tool That Turns DevOps Pipelines Into MLOps Pipelines

Really interesting post–TL;DR
This article explores KitOps, an open source project that bridges this gap by allowing you to leverage your existing DevOps pipelines for MLOps tasks through the use of ModelKits. A short walkthrough and code sample in the article will demonstrate how easy it is to get started.

https://thenewstack.io/kitops-is-the-open-source-tool-that-turns-devops-pipelines-into-mlops-pipelines/





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

Live Repository Execution System which can take args+file input - possible with GitHub Actions?

TLDR;

I want a users interact with a basic frontend where they can specify which repository they want to run, any program arguments to pass, and additional resources such as uploading files.
Based on this the repository stored in Github is run with the specified arguments + files, and output is returned to the user on the frontend.

Background

I am in the process of building a system which will allow my coworkers to easily run certain programs and scripts which are stored in Github, either to be executed immediately or set a scheduled time to run, via a basic Flask app. I want to be able to pass arguments, schedule runtimes, upload files, and receive output back on the frontend. I already use Github Actions to run scheduled jobs via cron, however, these don't take any arguments or file inputs, and don't require any returned files or outputs. I'm not sure with the current Github actions functionality if this is possible? I have thought about a system of spinning up docker containers which clones the necessary code and injects the user inputs, however, this seems like it would be much more of an undertaking. I'm trying to see if there is a way to do this currently that is fairly straightforward to set-up, either something existing in Github and I just build the frontend to send the necessary args/files, or if there is an alternative service(s) which would allow me to do this.

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

I want to understand how do guys manage your cloud infra & deployments at your companies

Hi Folks, could use your help, I want to understand how do guys manage your cloud infra & deployments at your companies
I am currently working on setting up the infra tooling just need your guidance

What type of infrastructure are you primarily using?
Kubernetes clusters
Standalone machines (EC2)


How do you manage deployments?
Configuration management tools (e.g., Ansible, Puppet, Chef)
Manual deployments
Github Actions

What Infrastructure as Code (IaC) tools do you use, if any?
Terraform
CloudFormation
Pulumi

Please go in as much details as you can
Want to understand what kind of setup is the most popular among the community

https://redd.it/1dngpx3
@r_devops

OpenTofu 1.8.0 Alpha1 is here!

https://opentofu.org/blog/help-us-test-opentofu-1-8-0-alpha1/

tl;dr: variables in module sources and backends. DRY up your configurations!

https://redd.it/1dnj0gh
@r_devops

How often do you work outside of your normal working hours?

Just wondering how often people here work outside of their normal working hours and for how long?

Say your working hours are 9-5 how often are you working before or after that time as requested by the company due to doing work which can only be completed when there are less customers using the platform?

I'm getting frustrated because it seems to be increasing and I value my time outside of work but I also understand what industry we are in.

Please let me know if there is a better subreddit for this question.

https://redd.it/1dnkaq5
@r_devops

I need some insight here...

The company I work for has me doing just about everything but DevOps. If I am not juggling developing ETL integrations with our iPaaS, I am managing the Tableau Server infrastructure, Dynamics 365, finance systems, etc.

When I do get to do DevOps tasks its... I get an email of a new user who needs access to a project, so I create a ticket and help desk creates the user and adds them to the AD group that adds them to that project.

A project needs a new app service - nope, not me. That's infrastructure so I have to create a ticket for someone over on that team to create the app service.

A project needs a service connection - nope, not me... that's again, infrastructure.

We have a new project - I have to create in Azure DevOps... then request help desk and infrastructure to create the AD groups, add the users, create resources and service connections...

What the eff am I even doing? Oh, hold up, I forgot the other thing I am expected to do... I'm also a project manager now? I have to have meetings with the project managers to ensure the projects are moving forward at appropriate speeds. I coordinate with business owners and the developers to ensure the project development is meeting the BO expectations... implementing features that are being requested etc.

I'm not doing anything DevOps related all. I have this effing certification that they made me get for this role and I'm not doing anything with it. Yet, my boss gets upset when projects aren't following policy. Ok, that's not my job though. I don't enforce policies, that's literally the job of the policy and security teams.

I get paid fairly well and shouldn't complain but I am because my boss wants me doing everything but DevOps and gets upset when I don't do the few things he wants me to do in DevOps on time... such as "standardizing devops projects to company policies".

I can't even develop PIPELINES! My boss wants the developers to manage their own pipelines and I am expected to just play support instead. The same developers who continuously cicumvent our branching policies by pushing their code to a different repository and then having one developer take that and push it into our repository.

Am I just being childish or something?

https://redd.it/1dnkmy0
@r_devops

Day to Day for a DevOps Engineer?

How do your days typically look as a DevOps Engineer (or equivalent)?

I'm a Software Engineer of 10 years with a deep interest in DevOps, and I'm a consultant so I often become "the devops guy" on whatever team I join. I'm curious how a dedicated DevOps Engineer role looks day-to-day to determine if I would like the switch.

https://redd.it/1dnn4h2
@r_devops

When I should be worried about idle time at work

I'm currently working as a cloud engineer in a data analysis team at a global automotive company. For the first four months, there was a lot of work and I was learning something new every day (I came from the SWE area), but for the last three weeks I've had almost nothing to do and there's no sign of any optimistic scenario.

Is it normal in this area (DevOps, Cloud) to have so much downtime? When should I worry? Sometimes I think they've hired wrong: there's no backlog and my leaders ask me to refactor SQL queries because they have no idea what else to ask for. I don't mind doing it, but I confess that it bores me and I feel like I'm stagnating.

https://redd.it/1dnr4r0
@r_devops

Automated telemetry capture using Python bytecode

https://jaywhy13.hashnode.dev/automated-telemetry-capture-via-python-bytecode-modification

This article covers my journey attempting to capture telemetry automatically using Python code. It ended up being super complex and too much for our company to maintain. I'm sharing it to get some insights to see if folks have done anything similar and have ideas!

https://redd.it/1dnu35e
@r_devops

Invalid login: 535 5.7.8 Error: authentication failed: authentication failed with Github Action

I try to send test report of my node.js project with github action automatically.

Here is the relevant part of github action

- name: Send test report email
uses: dawidd6/action-send-mail@v3
with:
serveraddress: smtp.protonmail.ch
serverport: 587
username: ${{ secrets.EMAILUSERNAME }}
password: ${{ secrets.EMAILPASSWORD }}
subject: Test Report
body: Test results attached.
to: <mailaddress>
from: ${{ secrets.EMAILUSERNAME }}
attachments: report.png


I always get the error shown below

`Invalid login: 535 5.7.8 Error: authentication failed: authentication failed`

`EMAIL_USERNAME` refers to my email of proton.me

`EMAIL_PASSWORD` refers to my email's password of proton.me

I also tried to use username of proton.me account but nothing changed. I still got the same error.

Where is the problem? How can I fix it?

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

Migrating to GitOps for the first time

Currently, I've been doing the CD part in this traditional way: deploy the chart using Helm and combine the values either from files such as values-dev.yaml, values-prod.yaml etc. These files contained the secret and non secret data, the non secret data are just empty strings and they get "injected" during the pipeline run from the secret libraries such as Azure Pipelines Library.

Pods have defined env vars or volumes and I inject the values into these from the values files using Helm syntax.


I am now looking to migrate this towards the GitOps world, so once again I will have the Helm chart defined in the repo, but this time I will be deploying it using ArgoCD installed on my k8 cluster where the Hem chart is also going to be deployed.

But I am now confused, how do I handle the secret values such as connection strings etc. which were before residing in the Azure Pipelines Library which was a part of pipeline. I can still use the files such as values-dev,yaml etc. because I can define them inside the ArgoCD CRD Application, but I can't inject the secret values into them anymore.


Am I doing this wrong, do I need to totally change the approach and ditch this way of using value files and secret data injection.? If yes, what would be the correct GitOps approach then?



https://redd.it/1do0p9f
@r_devops

Full IO manipulation Lib made for Lua

local dtw = require("luaDoTheWorld/luaDoTheWorld")


local concat_path  = false

local files,size = dtw.list_files_recursively("tests/target/test_dir",concat_path)

for i=1,size do
 local current = files[i]
 print(current)
end

Lib Link

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

Help with Linux server

I’m new to server setup and maintenance and am stumped trying to setup my Linux server for a personal project.

I’ve got a completely fresh install of Linux - only things I’ve done so far is setup SSH access, install nodeJS and npm and clone my GitHub repo.

I set my server running and saw the console log I’d set up to indicate it’s running fine on port 5000. However when I try to access it from google chrome or Postman, I get ERR_CONNECTION_TIMED_OUT and no console log from my application, which is set to console log a message whenever it’s hit.

I’ve tried setting up a simple HTTP server using Python and that’s not working either.

I’m sure I’ve missed something super basic so if someone could point me in the right direction I’d much appreciate it.

https://redd.it/1do2lzb
@r_devops

great o'reilly ebook deal on humble bundle

I've had many people recommend some of these books to me over the years, but it can obviously be expensive to try to build a library of them. Anyway, I don't even know how I ended up stumbling into this today, but I thought I'd share it. Considering that 90% of people in this sub are asking for advice on what to learn, I am certain that there is 25 dollars worth of value in here for most of you.

https://www.humblebundle.com/books/linux-for-seasoned-admins-oreilly-books

https://redd.it/1do3nm2
@r_devops

Observability - The eBPF Effect

I recently published the first of a two part feature on the impact of eBPF on observability tooling. The first part looked at eBPF in general, this part looks at how vendors such as Groundcover, Odigos, Beyla and others are using ePBF in their products as well as some of the challenges and limitations they face:

https://observability-360.com/article/ViewArticle?id=ebpf-in-observability-pt2

If you are using any of these in prod, it would be really interesting to hear about it.

https://redd.it/1do78g2
@r_devops

cool apps to deploy while doing labs?

im looking around for some lightweight apps that can be installed while doing labs strictly for learning at one's home setup. im feeling the usual "hello world" nginx/apache app is getting old. i know this is something of little importance im just trying to make learning and practicing more fun by having something cooler and fun as the end objective of the lab

https://redd.it/1do8nbu
@r_devops