How many of you actually do woodworking?

I've noticed it gets brought up often, pick up a hobby... Do woodworking. So my question is, how many of you DevOps folk actually do woodworking or is it just a jest answer?

I'll go first. I do. I'm in the middle of learning pocket holes to make a small bookcase

https://redd.it/10q1ovt
@r_devops

Is there a way to integrate a GitHub actions manual-approval step with a notification to a Slack channel?

Hey there,

I'm wondering if anyone is aware of a repo or example of a github actions manual-approval step triggering a message to a Slack channel (using a slack webhook url or even api token), which can then be clicked on to continue the pipeline?

​

thanks!

https://redd.it/10q2mnu
@r_devops

Kubernetes Workshop by Komodor - anyone joining?

Hey,
I saw that Komodor has a workshop today for Kubernetes and showcasing their platform, is anyone planning to join?
Have you tried it out?

https://redd.it/10q2m10
@r_devops

who do you follow?

Like title says, who do you follow on YouTube, Instagram, blogs.. etc.. To keep up with technology or learn quick new things?

I like to follow NetWork Chuck on youtube, I think he explains things very quickly and easy to understand.

So who do you like to follow on social media to learn, refresh, find out about new things?

https://redd.it/10q4mna
@r_devops

Noob Chef Attributes Question

in a default.rb file, what exactly do the value in the brackets stand for?

For instance,

default['apache'\]['app1'\]['version'\] = "8.1"

I have some idea, but I am still confused about how this is all organized and the general syntax.

https://redd.it/10q9ltf
@r_devops

Best CI system on Kubernetes

Hello all,
I’m looking for best CI system which can be placed on top of Kubernetes. I don’t have deployment part, however my team build and test a lot.
I have two main requirements:
1. Good integration with GitHub
2. Developer friendly

Now we have Jenkins(which seems to be too complex for my devs).
And a few GitHub Actions, however here I miss good self-hosting option, especially for aarch64.

My initial thought was Teton which isn’t very developer friendly for non-k8s developers.

Do you have any suggestions or opinions?

Thanks a lot!

https://redd.it/10q1xuk
@r_devops

Sync version of tool used in GitHub action and Makefile.

We use prettier.io in a GitHub action, and in a Makefile.

What's the best way to sync the version of both.

If we update the version in one place, and this update should effect the GitHub action and the version used in the Makefile.

In the Makefile we use npx with @2.8.3 at the moment.

https://redd.it/10q75bm
@r_devops

What are the best static analysis security testing tools for Terraform and infrastructure as code?

What are the best static analysis security testing tools? And is it worth going for a paid software rather than a free open-source one? I've heard of checkov, so I am wondering when you should use a paid software instead.

https://redd.it/10qfd0j
@r_devops

SycllaDB CDC connector

Hi Everyone,We are a new startup https://dview.io/ working in the Data Platform domain, currently at Pre-seed phase.As one of our first product we have created a new Syclladb CDC connector now available and open-sourced with :-

1. Checkpointing
2. Table level observability
3. UDT & Collection Type support
Sharing the Git Repo :https://github.com/dview-io/scylla-cdc

Say goodbye to data replication breakups!

Please share your valuable feedback which will help us to make it better.

https://redd.it/10q4et4
@r_devops

Auto scaling Self-Hosted solution for GitHub Actions?

GitHub Actions is surprisingly expensive. Now that they've released their runners with scaled CPUs, their prices are absurd.

This would be fixed by an autoscaling self-hosted solution. Does anyone know of something like this? This would be SIGNIFICANTLY more affordable.

https://redd.it/10qhid6
@r_devops

Question on workflows for GitHub actions

I’m new to Github actions (and DevOps in general) and I'm trying to make a basic workflow.

I have a single repo with sub-folders for the client and server. What I would like to do is set up a workflow to run some checks when I push code to the repo (linting, unit tests, etc.), but I can’t figure out the right way to structure this with GitHub actions. Should I create:

- Separate workflows for the server and client (e.g. server.yml, client.yml) with a single job for the checks.
- Separate workflows for the server and client (e.g. server.yml, client.yml) with a separate job for each check.
- A single workflow for the checks (e.g. ci.yml) with a job for the server checks and a job for the client checks.

Or am I overthinking it and all three are fine? How would you structure it?

https://redd.it/10q6jiy
@r_devops

SLI/SLO calculation via vmalert

https://victoriametrics.com/blog/rules-replay/

https://redd.it/10qkc9p
@r_devops

securing github actions from the inside

Having trouble finding info about this from searching, I imagine it's because I don't have the right search strings...

I need to implement a github action on our development repo (where devs play) that will dynamically retrieve secrets from Hashicorp Vault and do a little work on AWS on our behalf. That's easy. I can do that.

The thing that's rubbing me the wrong way is that anyone (ie devs) can effectively push a new branch to this repo that says 'on push to my branch LOLOLOL run this arbitrary workflow that retrieves the secrets from Hashicorp Vault and outputs them in plaintext, or hijacks the AWS session to do silly things.'

Okay, yes, first line of defense is the principle of least privilege - reduce the blast radius by using Vault policies and AWS IAM. But, like, I don't want a blast at all.

Is there a clever way to prevent developers from running arbitrary workflows? I'm envisioning some sort of "only run github actions from the main branch" or something, so the developer would at least have to get their workflow past a PR.

https://redd.it/10ql5y8
@r_devops

Preparing for the interview about GCP(Google Cloud Platform )

Hello folks! I'm a beginner DevOps engineer. How I can prepare for the interview about GCP? For which part do I need to more focus? In GCP what kind of questions, they can give me. Thank you!!!

https://redd.it/10qkb6n
@r_devops

Help needed Improve CI for multi repo java application

Hi everyone,
in my company I was asked to contribute to our process for CI. But I don’t really have a good idea.

We have a monolithic main software and it is using multiple layers of dependency in a final build.

Monolith -uses-> Module -uses-> Library


Monolith -implements-> Library

We use Jenkins to build artefacts that will be provided with JFrog for other builds.

We have jobs in Jenkins that are used when a merge request is created that executes automated test and after merging a publish job will be started.

Currently we are going to publish each change in the master of a module or lib manually and do some manual testing with the monolith.

I don’t think that only automating the job start of the current ones will be enough and I hope you can give me a suggestion that might be more a best practice.

Edit: format & typos

https://redd.it/10qjl18
@r_devops

Looking for a standalone GitOps controller?

I'm looking for an open source component that can pull changes from a Git repository, can be configured to trigger scripts when arbitrary changes are detected (e.g. config files, Terraform, et c) and can report progress and perhaps metrics. Ideally, it should be able to run as a container in e.g. Fargate or Azure Container Apps. This is essentially the Flux source controller as a standalone component and with no Kubernetes dependency. Does anyone know of such a tool?

https://redd.it/10qo9mf
@r_devops

Terraform / ansible

Hello people :) , do you guys know any free training website to learn terraform and ansible ? I just got my azure fundementals certification and need to boost my skill with terraform specially . Thank you

https://redd.it/10qp99h
@r_devops

many services, single host: what tooling to use? Ansible, terraform, docker, bash scripts?

Hi everyone,

I'm posting here to ask some advice because I don't see my particular setup being discussed often (if ever) in DevOps forums or articles. Some background: I'm a freelance web developer that builds a lot of different web apps for different clients, but that are each small enough to be able to host on a single server, that is: I'm hosting all apps I ever made on a single DigitalOcean host. Each service is contained within a Docker container and all networks, env variables and secrets are defined in several Ansible playbooks.

The problem is, with every service/app I add, my collection of Ansible playbooks becomes more a mess. There's dependencies between playbooks, for example: all services rely on a single Traefik reverse proxy (or nginx, I'm thinking of switching back to that) being started and my vars-files are starting to become a big list of variables. Aside from that, writing an Ansible playbook where I'm just starting containers feels like a bit of a pain compared to writing a docker-compose.yml and having a quick look at Terraform makes me think that this declarative approach is much easier to manage and reason about than the way Ansible does things.

Is there a better approach to doing this? Should I seriously look into Terraform? Kubernetes seems way overkill as I don't need any orchestration as the trafic I deal with is so minimal. Would it be better to just manually copy docker-compose files and nginx configs to my host and keep each apps config in their own repo? Should I try to make better use of Ansible roles and split out my configs in that way, defining common components like firewall and reverse proxy in a common role? Curious to hear how other people would approach this!

https://redd.it/10qpri2
@r_devops

What are the best devops stacks these days?

I'm working at a small company and I've been recently upping my skills especially in the cloud department and I'm pushing to go serverless, probably by using Google Cloud Run and AlloyDB. This has led me to question our use of Azure DevOps since it was here before I got hired and I never got to see alternatives.

I've heard that nowadays the best integrated solutions are either Gitlab with its CI/CD or GitHub Actions but it's unclear whether it's fully mature. My main issue with ADO is obviously the cluttered UI and the convoluted pipelines but that may be the nature of it. Also Microsoft which makes me more interested in Gitlab as a default but I don't want to choose a lesser solution out of spite for MS

Our DevOps is relatively simple, we manage projects using Agile, backlogs and the usual stuff. We have a CI pipeline that compiles everything, runs unit tests, publishes artifacts and then deploys on our on-prem windows server IIS instance for testing. Once it's validated main gets PRed into a release branch and after manual approved everything gets deployed to our production VPS.

In the future we will move for serverless and thus the pipeline will also build docker images and upload to GCR for production. For testing we will either retain an on-prem solution or find something else.

So I guess I'm looking for the pros & cons of each from actual user perspective in the context of our workflow

https://redd.it/10qq6om
@r_devops

How do you cope with your employer forcing you to use their shit ass laptop ?

It’s frustrating limiting, I can’t connect multiple monitors to it, and I can’t install any basic tool on it that are from my point of view absolutely necessary.

https://redd.it/10qoukj
@r_devops

´include:´ keywords limit (250) in gitlab ci pipeline

Hello DevOps community,

I just switched to DevOps and trying to move ahead in this exciting field. As part of my new tasks, I must look after a HUGE pipeline that creates GCP projects for the whole company using terraform as IaC tool. A week ago, it reached the include: keyword limit, despite the CI files are already DRY'd as much as possible; mainly do to the fact that the company is huge and there is a lot of demand for new GCP projects everyday, and for every project we create an individual ci file that performs terraform init, plan and apply. An idea that came to my mind is, to write a simple code (maybe in python or go), that on every pipeline run checks wether a file has changed or not, and depending on the changes generate a new ci file that triggers the whole init, plan and apply flow again, but it may take me too long to develop that code. Any suggestion on how to approach this issue is welcome and really appreciated!

https://redd.it/10qoyvs
@r_devops