System administrators around the world, we congratulate you on your professional holiday!
We hardly have system administrators left (well, almost). However, the legend about them is still fresh. In honor of the holiday, we have prepared this epic. Make yourselves comfortable, dear readers.
Once upon a time, the world of Dodo IS was on fire. In that dark time, the main challenge for our system administrators was to survive another day without crying.
Long ago, programmers wrote code slowly and in small amounts, deploying it to production only once a week. That's when problems occurred just once every seven days. But then they started writing more code and deploying it more often, leading to an increase in problems, sometimes everything fell apart, and rollbacks became more difficult. System administrators suffered but endured this chaos.
In the evenings, they sat at home with anxiety in their hearts. And each time, it happened: "never before, and again the monitoring system sends a signal for help: Dude, the world is on fire!" Then our sysadmins donned their red capes, put on their underwear over their tights, styled their hair into curls on their foreheads, and flew off to save the Dodo world.
Attention, a brief clarification. Classical system administrators who maintained hardware never existed at Dodo IS. We immediately had advanced cloud-based operations on Azure.
So what were they doing:
- if something broke, they made sure it got fixed;
- they juggled servers at an expert level;
- they were responsible for the virtual network in Azure;
- they took care of low-level things, like the interactions between components (*whispering* in which sometimes they had no clue);
- doing server reconnections;
- and many other wild tasks.
The life of the infrastructure engineering team (as we called our sysadmins) consisted of putting out fires and dealing with perpetually broken test environments. They lived and struggled, then decided to think: why is it so bad, maybe we can do better? For instance, why not stop dividing people into programmers and sysadmins?
Here's a task
Given: there's a sysadmin whose responsibilities include servers, the network that connects him to other servers, and infrastructure-level programs (web server hosting the application, database management system, etc.). And there’s a programmer whose responsibility is working code.
And then there are things that fall at the intersection. Whose responsibility is that?
Usually, this is where our sysadmins and programmers met, and it all began:
– Guys, nothing works, probably due to the infrastructure.
– Guys, no, it's the code that's the issue.
Once, at this moment, a fence began to grow between them, over which they joyfully threw poop. The task was tossed like poop from one side of the fence to the other. Meanwhile, nobody was getting closer to resolving the situation. Sad smile.
A ray of sunshine pierced the cloudy sky when a few years ago, Google came up with the idea of not tossing tasks around, but rather doing the common work together.
But what if we describe everything as code?
In 2016, Google released a little book called "Site Reliability Engineering" about transforming the role of the system administrator: from a wizard of magic to a formal engineering approach in software usage and automation. They went through all the thorns and obstacles themselves, learned from it, and decided to share it with the world. The book is available for free .
The book contains simple truths:
- doing everything as code is good;
- using an engineering approach is good;
- having good monitoring is good;
- not allowing a service to be released if it lacks clear logging and monitoring is also good.
Our Gleb read these practices (), and it took off. We're implementing! We are currently in a transitional phase. The SRE team is formed (there are 6 ready specialists, another 6 are going through onboarding) and ready to change the world, which is entirely made of code, for the better.
We are building our infrastructure in such a way as to allow developers to fully manage their environments independently and collaborate with SRE.
We predict instead of concluding.
System administrator is a respectable profession. However, alongside knowledge of the system side, excellent software engineering skills are also required.
Systems are becoming simpler and simpler, and super-unique knowledge in administering hardware servers is becoming less in demand with each passing year. Cloud technologies are displacing the need for this knowledge.
A good sysadmin in the near future will need to possess decent software engineering skills. Even better if they have strong skills in this area.
No one can predict the future until it arrives, but we believe that over time, there will be fewer and fewer companies willing to endlessly inflate their number of system administrators. Although, of course, there will always be enthusiasts. Just as today, few people ride horses; most prefer cars, yet the enthusiasts remain...
Happy Sysadmin Day to all, let the code flow!
Source: habr.com
