# The ONLY Roadmap to Get a DevOps Job in 2026 | Zero to Hero Guidance | Senior DevOps Engineer | USA

## Метаданные

- **Канал:** The Avi Narrative in USA
- **YouTube:** https://www.youtube.com/watch?v=9y3pcoUrTlA
- **Источник:** https://ekstraktznaniy.ru/video/30440

## Транскрипт

### Segment 1 (00:00 - 05:00) []

Hi, welcome to the Ravi Narative. I'm Ravi. I work as a senior DevOps engineer. Listen, I've been in the trenches since 2019. I started my journey at Clarivate, then moved through Prominence Insurance, Home Depot, PwC, Bank of America, Goldman Sachs, and now I'm back at Bank of America again. I've seen how the world's biggest companies like build their software. And today, I'm giving you the definitive 2026 DevOps roadmap. It worked best for me. I've guided to many people. Most people treat DevOps like a grocery list of tools. Going to give you a logic, like we're going to debug the career style from code errors to the DevOps bricks. So, this is my narrative, unfiltered, evolving, and unfinished. Let's get into the phase one. Linux, Linux. I keep suggesting Linux, Linux. Phase one is your grit. You cannot manage a cloud if you don't understand the OS. You need to master the Linux. Now, forget the GUI. Stay in the terminal now. Why? Because in collection, you're not just clicking any icons there. You need to know your system CTL commands to manage the services. And you need to like, you know, know how to SSH to get into the servers. You might need to like learn some grep and awk, and probably set to find that one error hidden like million lines of logs. Don't sleep on scripting. Whether it's bash for quick automation or Python for complex logic or AWS Lambda functions. Scripting, I feel like it's the glue of the DevOps, you know. If you can script it, you can scale it together. So, one I would suggest to like focus on not just Linux and bash scripting, but also like focus on file systems, permissions, vim, and lot more into shell scripting like different types of shells. One kind of certification I would suggest is to go with the Linux Essentials. There are some courses I would recommend. I'll put them in the description. Here's part two of the phase one, which I suggest like many of the beginners. Now, the chapter that everyone kind of skips when I keep talking is networking and storage. Why? Because 90% of the server down calls are actually the network or disk issues. You need to master of Like you need to master DNS records like, you know, A records or CNAMEs and routing, right? Routing is like a key concept in AWS. If your NACLs or security groups are blocked, your app is brick. Like, you know, you're just holding a brick there. You also need to understand the storage, specifically the NFS, the network file system. In a cluster, how do multiple pods like share the same data? That's NFS, right? So, also get comfortable with the YAML syntax. So, in DevOps, like, YAML is a language. So, one wrong indentation or one kind of a small misconfiguration can screw up the whole deployment. It would lead to a failure. One tip I would always suggest learn to use curl, you know, especially curl minus V and like dig and telnet. Like, if you can't trace a packet, you can't fix the system. So, you need to make sure that you get these commands in the line. like most familiar with you. So, I would say like networking and storage is like more important along with the Linux as a base foundation. Coming to the phase two, the cloud itself. People ask me, "Should I learn AWS, Azure, GCP? " You pick anything. Like, you know, it doesn't matter at the end of the day. Let's talk about that. Like, start with AWS, but don't get like vendor locked. You might ask me why? Because the concepts are resonant across every provider. Ask you bucket in AWS is just a blob storage in Azure or the cloud storage in GCP, you know, and EC2 is just a virtual machine. Whether you're working at a startup or a giant company, the solution is the same. Some companies might even run on-prem data centers. So, if you understand the architecture, like how a load balancer sits in front of a server, you can work in any environment. Stick to one cloud. Like, I would suggest to go with AWS because it occupies like almost 60 to 70% of the market now. So, stick to AWS. Learn more about IAM, VPCs, auto scaling, EC2s, EBS if possible, and like learn more about S3 buckets. Get to learn about EKS, which I'll come in the next topic. Of course, like, focus on lambda functions, too. Um, these are the most popular ones that I would always go with. Don't forget the networking part that I've said already. The best certifications that I would recommend is to go with the AWS developer associate. If you're good with that, go with the AWS solutions architect associate. These are the two certifications I would encourage to go with the DevOps. If you're a newbie in the AWS, once you have learned this AWS, you may ask me one question like, how do I manage the AWS resources? Now, we get into a concept of Terraform and Ansible. Why Terraform? Imagine you spent like weeks clicking through the console to set up an environment and someone accidentally deletes it if they have a right IAM access. If you have Terraform, you just run one command, it's back in minutes. So, that's like a kind of an insurance policy. You might ask me like, where do I store it? Like, you know, you might have to learn more about Git and Bitbucket. They're infrastructure as the code. Like, you know, sticking with Terraform is most important, to be honest. So, what's the purpose of the Ansible? Um why Ansible? So, if I have to update a config, or run a script on 100 servers, I'm not doing it manually. I'm using Ansible to push that change in 30 seconds. So, this is how you manage the server architectures at scale. You know, focus more on idempotency, playbooks, and some Terraform code and modules, if possible. Try to write your own modules at your own. So, that would make things more easier for you to learn quicker and pick up like any simple projects like, you know, and then and start writing Ansible playbooks, Terraform code, and how to like build AWS architectures. Phase four, this is my favorite part. Containers, Kubernetes. Oh, I might You

### Segment 2 (05:00 - 08:00) [5:00]

can call it Kube Kubernetes or Kubernetes, doesn't matter. Let's call it Kubernetes, by the way. So, but this is the big leagues. Docker and Kubernetes. Docker solves "It works on my machine" problem by packaging everything together. Kubernetes is the brain that manages it. So, it's kind of like an orchestration tool. Being a senior means knowing how to debug. Need to know how to troubleshoot a failing node. You need to check the Kubelet logs and you need to describe the pods to see whether they're stuck in the image pull back off or not, right? This is where the certifications come into picture. Now, this is the best way to get the certifications. Um you can start with CKA or CKAD and then get into CKS later. Um so, these certifications do actually matter in the DevOps market and it's not a it's not like a job guarantee, but they prefer people who have it. So, they force you to learn the commands that save you during an outage. Kubernetes knowledge also comes into the in uh Kubernetes as well. It's called as EKS and ECS like there are other services that use it as well. Try to focus more on the manifests and deployments. Try to get into some Helm charts that is super duper useful if you can if you have some time for that, but a lot of companies are already using the Helm charts for this. And trust me, if you know Kubernetes, you will know EKS, you will know AKS, and you'll know GKE, and you'll know OpenShift as well. I feel like it plays like 30% of the DevOps role. And try to learn Helm chart, it would make the deployments more easy for you in the Kubernetes. I'll put the link down below for more details. In phase 4. 2, this is what a lot of people talk about, CI/CD pipelines. For this, you don't have to go through depth Jenkins pipelines or GitHub Actions pipelines. Just stick to like one basic ones. Like you want to deploy something really simple like a Python application or Java application. Try to write a Jenkins pipeline on your own or like a GitHub Actions pipeline on your own. They're really free and like they're open source. You can start writing something in there. And that's how you build a CI/CD pipelines. There there's no gimmick about it. It's really easy. You just need to practice on it. That's all. Phase five, you built everything, right? But is it alive? No. You cannot manage what you don't monitor, right? You need Prometheus and Grafana for metrics. You want to see the heartbeat of your CPU and RAM. For log aggregation and searching in app logs, you know, you need the Splunk or the ELK stack. In enterprise environments like tools like DataDog and Dynatrace give you the full stack visibility. And don't forget alerting. If a service dies at 3:00 a. m., you need an automated alert to wake you up, you know? Or better yet, fix it, right? I also use a Rundeck for co-operational automation, giving self-service run books to other teams so they don't have to like call you every small task. There are many other tools out there. Try to stick to like one or two. You don't need to learn the whole thing. This should give you the coverage in alerting and monitoring aspect. I know this sounds like a lot, right? This is overwhelming, but these are the most common tools that cover you in every DevOps role globally. Once you have the foundation, the advanced stuff like, you know, GitOps like Argo CD or Spinnaker for multi-cloud uh our becomes really easy, right? DevOps is about continuous evaluation. Took me like 7 years from Clarivate to Bank of America to master this. So, take it like one phase at a time. Like, don't rush it. Now, once you have learned this, you'll be able to apply for SRE roles, platform engineering roles, or DevOps roles. You name it. Like, there are different terms for this. So, you'll be eligible for almost every single role if you have covered all of this. And the last thing I would say I'll tell you is like pick any 20 job descriptions that you wanted and you'll see like most of the ones that I've described. You know, you don't have to learn every single tool. Like, stick to one in each phase. The other tools or like other things will be like much easier for you. This is the Avi narrative. If you're ready to debug your life and career, hit that subscribe button. And I'm Avi, I'll see you in the terminal. Peace.