Learning through an AWS SAA course. Should I skip the Cloudformation section and learn Terraform instead?

I am currently learning through Cantrill's course to gain a solid understanding of AWS. However, I'm not completely sure whether I should get certified immediately or first aim to secure a job in order to gain practical experience using AWS at a production level before attempting the exam.

So I was wondering if it's a better idea to skip the cloudformation section and learn terraform instead, seeing as it's used in the industry more and is cloud-agnostic?

My two options (I'm in a time-constricted situation):

1. Learn AWS SAA, but replace Cloudformation content with Terraform. Then try to get on-the-job experience using AWS before taking exam. Won't have any cloudformation knowledge till I decide to take the exam and thus study for it.

2. Learn AWS SAA and learn Cloudformation. Get certified first, then attempt to find a job. I won't have learned Terraform.


Which option do you think is better?

https://redd.it/1417hzk
@r_devops

Azure/AWS spot machines

Hey, i want to know if someone here has managed to use spot vm in prod, is it even possible if you want to have 99% uptime?

https://redd.it/141a3pv
@r_devops

Building a Serverless Playground in Kubernetes using Knative & KinD

Hey fellow developers! I've just written a comprehensive blog post on how you can set up your own serverless playground in Kubernetes using Knative.

By following this blog, we will be going hands-on, building a serverless app from scratch in a local Kubernetes environment. The blog covers everything - from understanding serverless architectures, to the nitty-gritty of Knative and Kubernetes. It's a great resource whether you're just starting with serverless, or looking to deepen your understanding. Can't wait to hear your thoughts and answer your questions.

Check it out! https://karanjagtiani.medium.com/building-a-serverless-playground-in-kubernetes-using-knative-kind-88ddcbb4cca3

https://redd.it/141axig
@r_devops

Mimir Github actions metrics

What is currently the best way to collect metrics from GitHub about actions, such as duration, workflow status, error rate, etc? I found a few actions in marketplace, but they seem to be outdated. We are using Mimir with Grafan for metrics collection and dashboards

https://redd.it/141cwyj
@r_devops

Good monorepo tooling

Hey all

Wanted to ask the community here about any recommendations on goodmonorepo tooling. Have multiple different applications and helm charts running on k8s for context and was hoping to streamline our existing process. Not sure if theres something decent out there already or we just build our own

https://redd.it/141cgw7
@r_devops

The role of e2e testing and who manages it

I just joined this subreddit. I'm a fairly recent convert to DevOps. I had a nagging question... when it comes from the DevOps perspective, who manages end to end tests that span multiple departments? Each team has input and cares about the tests passing, but who manages the entire process to make sure they take place and coordinates with other teams.

If people have a great list of blogs or textbooks on this topic, I'd greatly appreciate it.

https://redd.it/141du5y
@r_devops

New CEO wants to understand DevOps impact on company bottom line. Help.

I work for a large apparel company (don’t want to name drop because I don’t want this to come back to me). We just got a new president/CEO who came up from the media side of things.


He spoke with the engineering heads and was asking for answers to questions the last guy never asked. Specifically, how to put $ numbers towards DevOps work.


Our previous head was usually just happy if nothing collapsed.


Has anyone gone through this? I'm worried he doesn't think the DevOps team is worth the money and wants to make cut.

https://redd.it/141i6xq
@r_devops

Out of Memory crashes in Jenkins pipeline

I run Jenkins inside of Kubernetes. The pipeline stages therefore run inside of a pod. I set the memory limit to 5 GB on the CI pod. When I poll kubectl top for the CI pod I never saw the pods memory usage exceed 1.2 GB. I also ran /use/bin/time -v on the crashing process to take a look at peak memory usage, which never even exceeds 1 GB. Still I’m experiencing out of memory crashes in the pipeline. What could be the cause of that? How can I debug that further?

https://redd.it/141hr0j
@r_devops

Gitlab Server Behind F5 Authentication

Hi! I have a Gitlab server that sits behind my F5. I want to be able to pull and push repos via HTTPS.
Right now from my local machine i would first have to hit the f5 and present my smart token/card to authenticate. From there traffic goes the a Virtual server > Gitlab.


Is there anyway for me to configure gitlab to prompt me to use my smart card in this process.
As of right now i am getting redirected once i hit the f5 and logged out due to the fact that im not using the cac via the CLI. Any other thoughts or suggestions are welcomed. Thanks.

https://redd.it/141l5dm
@r_devops

Short lived web-app

Hi,

I'm currently using github actions and whenever someone opens up a PR I want to deploy the web-app to AWS and for the deployment to be available for like ~1 hour. Currently we deploy the webapp in a docker container to ECS. Does anyone have ideas on how this could be accomplished?

https://redd.it/141his0
@r_devops

Which Kubernetes Certification is the Easiest One In The Land?

Looking for something to do while i'm off for a week next month? Which certfication will be the easist to earn

1. Certified Kubernetes Administrator (CKA)
2. Certified Kubernetes Application Developer (CKAD)
3. Red Hat OpenShift Developer
4. Red Hat OpenShift Administrator

https://redd.it/141t60v
@r_devops

Vendor lock in to AWS. Does going multi cloud make sense?

So we got a new CTO. Reasonable, sharp, technical guy. Coming from a big company which had on-premise servers.

He's been looking at our all AWS infrastructure, and recently asked aren't we been vendor locked in for AWS. And whether we should consider multi cloud.

I understand been multi-region for higher availability, but doing multi cloud just doesn't make sense to me, considering the implementation and maintenance cost and overhead. We're also a team of just 6 SRE engineers.

Does it make sense to go multi cloud for avoid vendor lock in? Is vendor lock in to AWS an actual problem? People who do multi cloud, what is the main reason for that?

https://redd.it/141uqit
@r_devops

New job opportunity

I’ve been a sys admin / sys engineer for the past 5-6 years and have been slowly building up certs to hopefully one day enter into the world of devops. Recently I got a job offer as a infra eng and I love all aspects of the job, but for config management they are heavy into vRA and saltstack. In my current job we are a heavy RHEL shop, and I even went out and got my RHCE for ansible. Is salt hard to learn? It looks similar enough to ansible, but I have no real hands on experience. I’ve heard from others that learning Salt would expand my toolset, and there are other aspects of the job that would help he grow professionally, but I am looking for any feedback for those who have used it. Is it pretty common in this world? How well does it integrate with cloud?

Thanks in advance!

https://redd.it/141wr0e
@r_devops

What are some of your favorite job search sites for Devops?

Looking for opportunity in new DevOps role after 5+ years with one company, wondering how to get started? focus on upskilling first? Which websites can help you learn new technologies and prepare for interviews specifically in the DevOps field? Are there any platforms dedicated to DevOps career opportunities?

https://redd.it/141unr8
@r_devops

Kubernetes confuses the heck out of me

Okay folks, I need someone to explain these items to me and how they are related or not in regards to getting a docker image to run in a kubernetes environment.

So I have the docker image working fine. When I run docker compose up -d, when I use docke desktop, I can see my desires output. Now I want to get it to Kubernetes.

What is a manifest? Is this a yaml file or just a word used to describe something else? How about deployment.yml and service.yml? Do I need both of them to deploy the image? Are they a requirement?

Also, how does Helm play a role in all of these? Does deployment and service belong to it? Do I even need it? How do all of these things play together?
What is a minimum required configuration to get a simple deployment going in Kubernetes?

All tutorials online are making a lot of assumptions about these terms and what they are and aren't.

https://redd.it/1420cib
@r_devops

Certs / courses for devops

Hello!
I have been doing "DevOps" for two years now and would like to know if there are any certs / courses I can/should do to boost my understanding of best practices. I am being asked to take course (s) certs from my work and attend conferences. Thing is, they don't know what I should attend and are asking me to pick something. I have googled a bunch but not entirely sure which one to commit to and I am open to recommendations and feeling a bit overwhelmed.

I would really like to take some courses that establishes DevOps best practices for end to end CI/CD and monitoring the application and how to best present the data as graph/tabular format for easy understanding for the end user (application developer).



Conferences planning to attend:
- kubekon (working out the price)

The tech I am familiar with:
- kubernetes (using it, not maintaining it)
- Prometheus
- grafana
- elk
- Jenkins (configuration as code)
- AWS (getting familiar with it)


Things I should be familiar with:
- I do not know what I do not know.

https://redd.it/141z8yf
@r_devops

Node Removal and Workload Migrations in Kubernetes

Kubernetes Node Drain -

https://faun.pub/a-deep-dive-into-node-draining-and-eviction-for-safe-node-removal-and-workload-migration-49a4c18d3211

https://redd.it/1422qef
@r_devops

Is there anything serverless cheaper than Lambda? I’m using it to process queued jobs, I don’t mind spot instances or guaranteed availability if it’s cheap

I run my queued jobs on Lambda serverless, so I can easily scale up or down depending on the load. I’ve processed massive queues in minutes but it’s not cheap.

I’ve seen Azure Functions and Google App Engine but their prices are comparable to Lambda. Is there something cheaper? Like going from EC2 to DigitalOcean VPS cheaper.

https://redd.it/141pg3s
@r_devops

Best sites to put your CV on

Hey all,

A couple years back I had a profile on talent.io and found this site good for attracting interest from companies. However I just tried to restart my account and it won’t let me. Perhaps they’re no longer in the UK or no longer do DevOps roles.

Are there any sites people recommend where you can put on your CV/profile and companies come to you?

https://redd.it/141fyju
@r_devops

Introducing Selefra: Open-Source Policy-as-Code Software for Multi-Cloud and SaaS Analytics

I'm excited to share this article with the DevOps community. In this article, we'll introduce our product, Selefra, which provides an open-source policy-as-code software that offers analytics for multi-cloud and SaaS environments.



Terraform, as you know, provides a platform-agnostic approach to infrastructure provisioning and management. It allows you to define, create, and manage infrastructure resources across various cloud providers, on-premises data centers, and other infrastructure platforms. It supports major cloud providers such as AWS, Azure, GCP, and others, along with infrastructure platforms like VMware, OpenStack, and Kubernetes. This flexibility enables organizations to adopt a multi-cloud or hybrid cloud strategy while ensuring consistent management across different environments.



Selefra specifically focuses on the multi-cloud landscape. With Selefra:
\- You can quickly build a data lake across cloud assets, manage security configurations, and analyze resource consumption.
\- It allows you to write SQL and YAML to develop Infrastructure as Code (IaC) programs for tasks such as misconfiguration detection, cost analysis, security provisioning, compliance scanning, and more. Our solution follows a declarative approach, empowering you to efficiently define policies and rules.
\- Selefra Cloud offers out-of-the-box modules, including CIS Benchmarks, SOC, GDPR, HIPAA, ISO, and more in [Selefra Cloud\](https://app.selefra.io/). These modules provide pre-configured policies and compliance checks, making it easier to ensure your infrastructure meets regulatory requirements and industry standards.



To interact with various cloud providers, SaaS providers, and development platforms, Selefra relies on provider plugins. These plugins establish connections and enable seamless integration with different services and platforms.



We deeply value the contributions from the Terraform community, and it has been a great source of inspiration for us. We welcome any suggestions and questions you may have. Your feedback and ideas will help us improve and enhance Selefra further.



We highly encourage you to explore Selefra and embrace the power of GPT for accelerated and efficient cloud security analysis.



Website: [https://www.selefra.io/\](https://www.selefra.io/**)
GitHub: [https://github.com/selefra/selefra\](https://github.com/selefra/selefra**)
Twitter: [https://twitter.com/SelefraCorp\](https://twitter.com/SelefraCorp\*\*)



Thank you for your support!

https://redd.it/1427dry
@r_devops

Understanding AWS NAT expenses

Hello folks, on mobile so please forgive my formatting.

Due to recent infrastructure changes my company has seen a drastic spike in NAT gateway outbound traffic expenses. This was a surprise expense to us that (to be frank) we should’ve predicted.


We use AWS lambdas in private subnets spread across 3 AZs, each with their own NAT. Lambdas are called via API Gateway. We understand that traffic outbound from private subnets exits via the NATs, but since this change we are bleeding $500+ a day.

We have researched a number of different options which we have started testing, but I wanted to reach out here to see if anyone has encountered a similar issue. Majority of our traffic is HTTPS bound out the NAT.

Do we have any good options to bring our traffic down or redirect in any way?

Best!

https://redd.it/141dajc
@r_devops