Questions:
- What is devops methodology, the role in the production of software products, what is the complexity of the search.
- Types of professionals applying the devops methodology
- Where is there to be, DevOps-engineers / SRE went and came to the market
- Do you need a DevOps Engineer / SRE? If so, which one?
- Search channels
- What to look for in a resume
- How to start a dialogue
- We will call you back - why is it impossible and what does it lead to in the field of search for devops
1. What is the Devops methodology, the role in the production of software products
Devops is an acronym for development and operations - software development and operation.
DevOps is, first of all, a philosophy and methodology for improving engineering and development culture, not only within one team or project, but within the entire company, because DevOps implementation introduces changes not only in the development process, but also in the company's business processes.
The role of this methodology in software production: simplifying processes, avoiding errors, establishing communications, control, monitoring and logging, security control.
Applying devops methodology can be compared to the cement that holds the bricks of code, processes and outputs together, or to a pipeline that accelerates the development process, fixing bugs and delivering new functionality.
2. Types of professionals applying the devops methodology
Who works with the devops methodology? The entire development team as a whole. Testers, admins, developers, information security specialists It's like agile / ITSM / ITIL, only DevOps.
Simplified development pipeline (pipeline): Code is written (dev) - combined if written by several programmers (merge) - tested (test) - sent to assembly (build) - production :)
That is, all it specialists at all stages use devops methodology and tools:
CI / CD - tools for continuous integration of pieces of code with each other and delivery of code to where it is required: packages, containers, etc. The final view of the application.
CI part:
Development - code development and analysis, its parts:
Git - version control tools, code merging. First, the code is merged within one repository, and then it is built and then tested;
Build - assembly;
Test - continuous testing tools that provide feedback on business risks;
CD part:
Release + Deploy - change management, release approval, release automation;
Configuration - Infrastructure Configuration and Management, Infrastructure as Code;
Monitoring - monitoring application performance, end user experience.
What for what, a short guide
// so that you do not rub the game in the interview:
for building infrastructure - Terraform or utilities of the cloud provider
of the configuration management system - Ansible, Chef, Salt, Puppet
common CI / CD tools - GitLabCI, GitHub Actions, Jenkins, TeamCity etc.
for containerization - Docker, Kubernetes, Nomad, OpenStack, etc.
When you know what docker is for, they will communicate with you in a completely different tone.
3. Where there is to be, Devops engineers went and came to the market
Yes, we will call them that. Because they have nowhere to go, everything else is very cumbersome, poorly understood and unmemorable. They will have to come to terms with the fact that recruiters call and will call vacancies "Signor Devops".
Basically, there are 3 sources from where those that we hunt for the request “Devops urgently needed” come to the market:
The first and most numerous group: former and current system administrators. It's easiest for them: they mastered additional. tools and done.
Second group: developers who decided to go into devops practice. There are fewer of them, they need to master Linux and administration.
The third group: “I woke up and realized that it was mine” - the guys who took the courses “Devops in 3 weeks” or something more intelligible. I have no purpose to advertise, so you can google it yourself. With these, from the point of view of recruiting, work is both easier and more difficult. On the one hand, they are taught modern and popular instruments. On the other hand, there is no experience and understanding either in administration or in development. A good plus for management: their salary expectations are lower than those of the first and second groups.
June-middle-signor
June - knows how to maintain already implemented tools, but cannot implement from 0. Needs mentoring. At the same time, he can be a good admin.
Middle does not have a deep understanding of the methodology - he can introduce some tools from 0 and teach others.
Middle must have a good understanding of the methodology, understand the practices, RnD, can independently choose the tool to use
Signor - can put all devops practices from 0. To defend architectural solutions. He understands the risks for software development, chooses all the tools himself. Arguably proves his choice.
4. Do you need devops / SRE? If so, which one?
If the vacancy customer is a product team with internal development - yes, you need it.
If it's just a code for outsourcing, then it's not necessary.
What kind of devops your team needs: depends on the product.
Most often, you need Linux admins with experience in writing scripts in popular languages.
If specific software is being developed, devops must understand the nuances of this development and it is worth looking for those who have moved into this area from development on a similar stack.
And a little about SRE:
Site Reliability Engineering is practically the same as devops, if you do not go into the details. But we are not engineers and we will not go deeper.
SRE is a set of methods, indicators and prescriptive ways to ensure the reliability of systems. The word "site" in this context is read as "system" or "platform", and not a website in the usual way. SRE - ensuring the reliability of all levels of the system: from physical to logical, which means that SRE is a kind of conglomerate of a developer (yes, SRE should be able to code) and a system administrator with all that it implies.
SRE is a kind of offshoot, or rather, its own implementation of the DevOps direction from Google.
5. Search channels
The main channel for searching devops: the DevOps Jobs telegram channel - work and analytics.
Habr and the link show themselves well, a little worse than FB and generally not suitable for searching for HH.ru and SuperJob, while decent admins are quite looking for there.
Difference from the search for developers: the devops community is very friendly and sociable))) If a vacancy is published in such a way that it only causes laughter or facepalm, be sure that it is already being discussed in the telegram.
6. What to look for in a resume
We all know that a recruiter evaluates a resume in 3-5 seconds.
In addition to the general rules for evaluating resume, which you already know:
Should be: GitLab, GitLab CI, Ansible, Docker, Terraform, Zabbix, KVM, MySQL and PostgreSQL, Prometheus, Grafana, ELK stack, Jenkins, K8S / Kubernetes, AWS \ Azure \ GCP \ Yandex cloud \ Mail Cloud.
This is devops.
There is something of this and the words Windows 7 \ 8 \ 10 \ Server 2012 \ Server 2016 and so on. - the former Windows admin.
Cloud technologies
If you see the word Azure - this is a cloud from Windows
Everything else: GCP, AWS, etc. - these are clouds in which Linux systems and their majority prevail.
There is a phrase: I taught in courses GitLab, GitLabCI, Ansible, Docker, Terraform, Zabbix, KVM, MySQL and PostgreSQL, Prometheus, Grafana, ELK stack is a student.
Not everyone works with clouds. A devops that does not work with clouds is a devops that works in a closed circuit, a data center, a DC, etc. He needs to develop =) Cloud technologies are the future.
7. How to start a dialogue
Very simple. Good evening, looking for a devops. Here's the description, here's the plug, here's the conditions. Waiting for an answer like the nightingale of summer.
The vacancy MUST contain:
Fork. Fork - 2 understandable numbers. From 0 to 800k is not a fork, this is nonsense.
Conditions: office / remote, what else is additional: project \ partial \ full-time
Description of the development stack. It is important.
Description of tasks. Support for an existing one and implementation from 0 are very different things. If you have an architect, this is one thing (he just needs hands), and if he does not, then you also need a specialist who knows how to architecture, and not just a blooper and you
're done Life hacks : read the early bash.org (until it became bash.im), watch anime, play games, be adequate, do not disappear without feedback and you will be immediately noticed.
8. We will call you back - why it is impossible and what does it lead to in the search for devops
As we have already said: the community is friendly and will discuss you right away. It's easier not to lose your reputation than to restore it.
It doesn't matter who you hunt: YOU MUST GIVE FEEDBACK. Even sad. Formulate it appropriately. Better a bad ending than waiting without end.