We explain what skills a DevOps engineer should have, what tools they need to be able to work with, and where to learn all of this.
How should DevOps training be structured, what skills should a qualified specialist have, and what technologies will they need to work with? We discuss this in our article.
Who is a DevOps engineer and what do they do?
A DevOps engineer is a specialist who supports and automates the software development process at all stages: from the beginning of development to deployment and delivery of the product to users. This means that they:
- prepare the infrastructure for development and testing, install the necessary software on servers, and configure it;
- creates and maintains CI/CD — processes for automatic testing and code deployment;
- configures monitoring to collect all the necessary metrics on system performance.
We have put together a DevOps engineer training plan that will allow you to master the Development Operations methodology from scratch.
Where to start learning DevOps
You need to start with the basics: operating systems, network technologies, virtualization, and scripting languages. Let’s take a closer look at each of these points.
Computer networks
As a future DevOps engineer, you will have to configure the interaction of various systems and services with each other. As a rule, most of these interactions take place over the network.
A DevOps engineer does not lay network cables or configure routers. However, they must understand the general principles of how networks work and are built, as well as the differences between network layers in the OSI model — this will allow them to quickly find the source of potential problems. If a program written by developers cannot access the database, your task is to analyze the situation and figure out whose side the problem is on.
Operating systems
Linux is an important part of DevOps training, as most often such specialists work with Linux or other *nix systems, but sometimes Windows Server can also be encountered. Therefore, you need to learn how to work in different operating systems, understand their differences and features, and know the principles of operating systems: what an OS kernel is, how the system works with memory and allocates it to processes, how file systems are structured, and so on.
This knowledge will help you understand the architecture of the systems and tools that you will study further. It will also help you solve problems that will inevitably arise: a service won’t start, a service returns an error, etc.
You also need to be able to work in the console (especially for *nix systems) and know the basic commands and utilities. For *nix systems, you will be working with the Bash command shell, and for Windows, with PowerShell.
Python
Python is one of the main programming languages used by developers. It is a simple and convenient language with a low entry threshold for writing automation scripts. Ruby or Go are used much less frequently for this purpose.
DevOps specialists do not need to train to the level of a senior developer and know all the intricacies of programming. But in order to automate processes competently, they must be confident in writing in their chosen language.
Virtualization
Nowadays, it is difficult to find an infrastructure that does not use virtualization. Most often, several virtual servers run on one physical server. And since a DevOps engineer prepares the infrastructure for development, testing, and code deployment, they need to understand virtualization technology, be able to create virtual machines, and know the specifics of working with them.
Learning DevOps: advanced level
Now let’s take a look at the DevOps tools that you will use to solve the tasks assigned to you.
Configuration management systems
DevOps engineers prepare servers for operation: they configure them and install software. But they don’t do this manually, they use automated configuration management systems such as Ansible. Most often, configurations are described using YAML, a data serialization language.
CI/CD processes
One of the main tasks of a DevOps specialist is to create and maintain CI/CD processes that automate code assembly and deployment. Programmers write code, and DevOps engineers ensure that this code is automatically tested, assembled into a finished program, and deployed to the production server if necessary.
Ready-made tools such as Jenkins or Gitlab CI are used for this purpose. A DevOps specialist must be familiar with all stages of software development, understand which systems are involved, and know how to automate the entire process.
Cloud technologies
This is a very important part of the DevOps training program. Many companies use cloud platforms to host their infrastructure, and some of them use a multi-cloud approach, which involves using several cloud platforms at once.
DevOps is responsible for the cloud infrastructure. They create and configure virtual machines, create database clusters, connect other necessary services, and link them together. Therefore, they must understand the specifics of working in the cloud, be able to build cloud-native architecture, and understand different types of cloud services. And if the company has its entire infrastructure in the cloud, then such a specialist can essentially replace the system administrator.
Microservice architecture and containerization
Another popular trend is microservice architecture and containers. Many companies develop their software so that it wraps and runs in Docker containers. If there are too many containers, an orchestrator such as Kubernetes comes into play to help automate their management.
And while a developer can still write a Docker file and wrap their application in a container, a DevOps engineer will be responsible for deploying containers and managing Kubernetes.
Monitoring
Even after the servers are created and configured, and the applications “reach” production in automatic mode, the work of DevOps does not end there. Now all this infrastructure needs to be monitored: monitoring needs to be configured, metrics need to be collected, and graphs need to be visualized.
Ready-made tools such as Prometheus and Grafana are used for this. You must be able to install and configure them, as well as understand which metrics to collect and what to pay attention to.