دانلود کتاب ایجاد بنیادهای SRE – راهنمای گام به گام برای معرفی مهندسی قابلیت اطمینان سایت در سازمان های ارائه دهنده نرم افزار

This book is based on a site reliability engineering (SRE) transformation journey from a real software delivery organization in the healthcare industry. The organization runs a cloud-based platform for medical applications and services. The platform is deployed in many data centers and the applications on the platform are used in hospitals around the world. Some of the applications are used when patients are in critical condition. It follows, then, that the platform’s reliability is of paramount importance. But what is reliability? How do you measure it? How do you create an environment where development teams are motivated to invest in reliability? These were the questions I grappled with several years ago when the organization struggled to provide a reliable platform for applications and users alike. High-profile customer escalations were common. People were unaligned regarding backlog prioritization of new features versus reliability work. The operations teams were struggling to operate the product. The development teams happily implemented new features but paid very little attention to how the existing features were running in production. Project management plans were impacted greatly by deployment of large numbers of unexpected hotfixes. High-profile customers called the leadership team demanding that the service be restored or that missing features be delivered. Everyone had an opinion on what needed to be done to improve the situation, until the next outage took place, causing new opinions to emerge. I had attended the QCon London conference for several years. The conference helped me stay abreast of new trends in software development and operations. SRE was one of the topics at the conference. I was aware of its existence but had not started learning about it. At one of the QCons, an entire track was devoted to SRE. I spent a significant amount of time attending sessions in that track. At the end of the conference, it was clear to me that SRE was gaining momentum in the industry. While traveling back to work from the conference and looking over my notes, I decided that it was the right time for the organization to try SRE in an attempt to improve operations. There was no other structured approach to doing operations that I had come across. What we tried ourselves without SRE did not yield visible improvements. Many companies at the conference reported being successful, whatever that meant, in doing operations using SRE. Getting started seemed to be easy. It would only take a couple of basic indicators, like availability and latency, the definition of acceptable targets for each service, and alerts on when the targets would be broken.