Schema migrations on RDS from github actions

We run a number of laravel apps that require schema migrations on deployment from time to time. These RDS instances are in private subnets so obviously github actions can't reach them.

Our current solution is for the GH workflow to trigger a jenkins task from a jenkins machine we have in the same vpc.

Not particularhappy with this solution (although it works fine) and wondering if anyone has any other suggestions or ideas on how to approach this

https://redd.it/xz4gvm
@r_devops

DevSecOps Azure

Goodnight guys! I just had a 1:1 with my manager and we talked about career progression since I’m in my early career. I work at a consulting firm and you can play many roles. I work under the azure security team and wanna do DevOps. In the future I would like to be a solutions architect. He gave me some topics I should like into and get comfortable with Azure Bicep & DevOps, Terraform, and Python. He told me I can use Python as a wrapper for Terraform, I’ve been Googling and I don’t know what exactly that’s means any suggestions? Links? Videos? Guides? On how to do that. Was also thinking about doing that coursera Google python automation course to get my python skills up.

https://redd.it/xz88j3
@r_devops

Great resource

https://github.com/bregman-arie/devops-exercises

https://redd.it/xz9sa7
@r_devops

How do you separate your vpc network?

What are the factors you're looking at for a microservice application? Below are some of the things that comes into mind:

* Each application has its own VPC network
* Everything resides in the same VPC network but each application is separated using different sets of subnets
* ...

https://redd.it/xzc9dz
@r_devops

Please suggest infrastructure for a SaaS app

Hello,

I'm here looking for help/suggestions after lurking around a lot.

I've created a web app which consists of the following components (docker containers):

|App|CPU (limit)|Memory (limit)|
|:-|:-|:-|
|backend PROD|250m|512Mi|
|2 x workers PROD|250m|256Mi|
|API PROD|250m|512Mi|
|backend DEMO + API DEMO|250m|256Mi|
|2 x worker DEMO|250m|128Mi|
|3 x JS apps (js, SSR, static)|200m|256Mi|
|helper|200m|256Mi|
||||

Right now they are all hosted on a Linode kubernetes cluster connector to DigitalOcean managed DB.

Why? I wanted to learn a little bit of kubernetes and needed a secure place to store my data (with backups and point in time recovery if needed).

But I'm looking for alternatives, the cost of all this is almost $80 per mo (cluster, db, load balancer).

I need a place where I can just throw a dockerfile and in return I get a domain.

I've managed to move JS apps to Vercel hobby, which is more than enough for me. But I still have the other components.

What would you do to lower the cost of this and also managed less infrastructure? (I don't wanna deal with VPS upgrades, hardening, patches...etc)

​

PS. currently looking fly.io , they have some cheap VMs I can use for workers ($2/mo) and even some free ones. I'll make some calculations.


LE: JS apps were moved to vercel

​

Thank you.

https://redd.it/xxucvm
@r_devops

Choosing the most convenient software for my business

Hello there. I just started testing the Azure DevOps platform. It looks great but from experienced users in these platforms, how to decide which one would suit my business more? Also, what things should I consider when I choose a DevOps platform? I use Azure Cloud for my work and GitHub, Gitlab, Kubernetes, and Docker. I tried Jira and ClickUp apart from Azure DevOps (a longer time than AzureDevops). I want to establish my startup in one single DevOps platform.

https://redd.it/xxtx8r
@r_devops

How to do 1:1:2 traffic loadbalancing for 3 apps in K8s with ingress?

Requirement - Deploy 3 applications each with a traffic weight - 25%,25%,50%.

I was able to do a canary load balance with ingress for 2 apps with 25% and 75% by creating a simple ingress for old-app and a ingress for new-app with canary-weight == 25%.

But I'm unable to figure out how to do it among three apps.


Note - this has to be done withing k8s,minikube. Three apps will be deployed with their own services, then ingress layer to be added.
Any perspectives, hints, or solutions are welcomed.

https://redd.it/xzgk4e
@r_devops

Q/A opensource tool for organisation wide?

Hi there, I wanted to onboard my company to use an org-wide common tool for asking questions/ans. Basically a forum to discuss anything.Are there any recommendations for such a tool? Or any better way to organise this?

Note: I am not looking for tools like slido, something that can be used as a question bank.

Below is an example of one paid tool

https://www.discourse.org/

https://redd.it/xxn8p6
@r_devops

How do you work on personal projects when services are so expensive?

I’m thinking of getting into devops but something that deters me from becoming invested is the cost/risk associated with everything you do in the cloud. Compared to general scripting/programming of small projects run locally for no cost, what projects do you like to make in cloud service providers & how do you not get bothered by the costs associated with them? I’m aware of freemium plans, and I’ve only worked a little in AWS, but the possibility of accidentally choosing a wrong plan and getting a steep bill are worrying. I guess my main question, coming from a noob is what personal project ideas interest you in the cloud for practice do you like to work on that don’t burn a whole in your pocket? Thanks

https://redd.it/xzk63q
@r_devops

New job + Anxiety help?

Hello there, let me explain. I have like 2.5y of experience in DevOps.

Started this year at 20K, then jump to another company for 41K, and now just did the same for 74K. Let me tell you this is a lot of money at least in my area.

I've been suffering from anxiety from day 1, thinking about im not the right for this job, that even if I have passed through 5 tech interviews 1h each, I'm a fraud and I'm going to get fired soon. I feel like bullshit and seems that everyone else here have so much experience from me.

Have you ever felt something similar? What is something I could do to deal or handle it?

https://redd.it/xzo0zy
@r_devops

Moving from DevOps to SWE role

I interviewed last week internally for a Java Server Dev position and although they are still interviewing externally, the hiring manager and another senior dev I work with in that department told me they would like to move forward with me if I’m interested. My current position is devops engineer. This Java role has no on call. Has anyone here moved from a devops role to an swe role? How was the transition?

https://redd.it/xznsaq
@r_devops

What is the difference between the term api testing and interface testing?

Aren't they both testing how data is transferred from the mating software?

https://redd.it/xzqfkg
@r_devops

what questions to ask to the interviewer?

Hello,

I started giving interviews and I am always hesitating when the interviewer asks me if I have any questions. Then I have to think twice what to ask and I ask this:

"I like to keep myself updated with the current industrial trends. Does your company has any employee education program or support for their certification?"

I am not sure if that question is a good one. Could you please share what would be the good questions to ask?

I was told that it is must to ask questions at the end of the interview otherwise it shows that you are not really much into learning about the company. Not sure how true is that.

Thank you.

https://redd.it/xzrwfb
@r_devops

How do you distinguish between end to end and integration testing?

I've seen them used interchangeably a lot, and I've seen projects that have both without a clear distinction for what the difference is.

I've ended up just dropping the phrase integration test from my projects and only using e2e, as I've found it to be less ambiguous.

Is there a real definition out there somewhere that draws defining lines for when a test would go into one suite or the other? Or is there just naturally a lot of overlap? And if there is a lot of overlap, does it make sense for one project to have both types of test suites?

https://redd.it/xztk9a
@r_devops

Concourse task: Getting log on erro?

I have a ton of concourse jobs that automate deployments. If a task fails, I post to a special slack channel via an email with just a note that it failed. I'd like to attach the log for that task so I can see what's going on without having to log into concourse.

onfailure:
put: send-an-email
attempts: 1
params:
subjecttext: "((ENVIRO)): Failed to create BTP stemcell pipeline"
bodytext: Check concourse pipeline at <<path to pipeline>/${BUILDPIPELINENAME}/jobs/${BUILDJOBNAME}/builds/${BUILDNAME}"

I'm not a Concourse pro. Can someone point me in the direction how to get the task log and attach it to the email?

https://redd.it/xzxkmk
@r_devops

Getting build log on failed Concourse Task?

I have a ton of concourse jobs that automate deployments. If a task fails, I post to a special slack channel via an email with just a note that it failed. I'd like to attach the log for that task so I can see what's going on without having to log into concourse.

onfailure:
put: send-an-email
attempts: 1
params:
subjecttext: "Pipeline failed"
bodytext: Check concourse pipeline at <<path to pipeline>/${BUILDPIPELINENAME}/jobs/${BUILDJOBNAME}/builds/${BUILDNAME}"

I'm not a Concourse pro. Can someone point me in the direction how to get the task log and attach it to the email?

https://redd.it/xzygqd
@r_devops

thoughts on book: The unicorn project by Gene kim

Have read the pheonox project and it was decent read. What are your thoughts on it's sequel

The unicorn project

https://redd.it/xzun9w
@r_devops

Cloud computing adoption survey: University research

Hi all,

I am an information systems management student, currently writing my dissertation about cloud computing adoption for European and US SMEs.

The goal of the dissertation is to potentially create a new cloud onboarding strategy that companies can use when deciding to move their operations from on-premise to the cloud.

Part of the research consists of investigating the organization's experience through the use of an online survey.

If you have been part of a digital transformation that included the adoption of cloud services I would really appreciate if you could spare 15 minutes to share your insights in the following Google Forms Survey: https://forms.gle/9cdNFwJX3DVHCuYG8

Thank you.

https://redd.it/xzu6z8
@r_devops

Build system for a multi repo project

I'm not a DevOps engineer but I take on a lot of DevOps and pipeline building tasks at work. I work at a start up so neither my team nor I have a lot of experience with build tools or organizing growing projects into something that isn't a convoluted mess. Since I don't have anyone I can ask what the best way to do this, I wanted to ask people with more experience how to handle this growing ecosystem.

&#x200B;

Problem:

We have 5 projects that are all distributed across 5 different repos with their own dependencies and setup scripts. They are all developed separately and when you string all of them together you get a full application. The issue is that development and testing the system as a whole or the integration of a couple of systems requires going into each repo, checking out the correct branch, reinstalling changing dependencies, deploying it on the local machine and then working with it. There are a lot of automatable steps but they all have to be done manually. The majority of these repos have been containerized.

&#x200B;

Solution 1:

What I want to do is have a build system that you can feed a couple of parameters to and it will spin up the different components, deploy them and provide you with an application ready to test. My first pass was to throw a couple of the components into a docker compose file and deploy it. This doesn't scale well since I need to be able to deploy any of the 5 projects in various combinations so I would need like 5\^5 different combination of docker compose files.

&#x200B;

Solution 2:

My second plan (where I am right now) is to write a command-line python program that you provide args for which projects you want to pull and deploy and create a dynamic docker-compose yml file. As I was planning this out, I realized I was just trying to build a custom build system mostly as a wrapper around docker compose. I'm not opposed to that but if a tool to do this already exists I should leverage that rather than try to reinvent the wheel. Eventually I want to put this into jenkins as well to make the CI pipeline better for integration level testing.

&#x200B;

Advise?:

So I wanted to ask, is there a tool that is better suited for this? Since this is all local I didn't think ansible was the right tool. I looked into bazel and that didn't seem like it was the right tool either. The issue isn't compiling the different projects, its deploying them in a custom manner. I'm open to any critic of my thought process right now.

https://redd.it/y05xdi
@r_devops

Is deploying a Nomad config the same as deploying back end code to a server?

I am guessing it's just ssh into the server and then calling git pull and then running the orchestrator with "nomad run" in the shell, but I am wondering if it's actually more complicated.

https://redd.it/y06ua3
@r_devops

What is the most esoteric database found in your production environment?

I want to learn more about other databases, see if there’s one that covers an interesting use-case.

https://redd.it/y01yke
@r_devops