Book Image

Building Microservices with Go

By : Nic Jackson
Book Image

Building Microservices with Go

By: Nic Jackson

Overview of this book

Microservice architecture is sweeping the world as the de facto pattern to build web-based applications. Golang is a language particularly well suited to building them. Its strong community, encouragement of idiomatic style, and statically-linked binary artifacts make integrating it with other technologies and managing microservices at scale consistent and intuitive. This book will teach you the common patterns and practices, showing you how to apply these using the Go programming language. It will teach you the fundamental concepts of architectural design and RESTful communication, and show you patterns that provide manageable code that is supportable in development and at scale in production. We will provide you with examples on how to put these concepts and patterns into practice with Go. Whether you are planning a new application or working in an existing monolith, this book will explain and illustrate with practical examples how teams of all sizes can start solving problems with microservices. It will help you understand Docker and Docker-Compose and how it can be used to isolate microservice dependencies and build environments. We finish off by showing you various techniques to monitor, test, and secure your microservices. By the end, you will know the benefits of system resilience of a microservice and the advantages of Go stack.
Table of Contents (18 chapters)
Title Page
Credits
About the Author
About the Reviewers
www.PacktPub.com
Customer Feedback
Preface
Index

Patterns


The truth about microservices is that they are not hard you only need to understand the core software architectural patterns which will help you succeed. In this section, we are going to take a look at some of these patterns and how we can implement them in Go.

Event processing

In our case study, we failed due to a downstream synchronous process failing, and that blocked the upstream. The first question we should ask ourselves is "Does this call need to be synchronous?" In the case of sending an e-mail, the answer is almost always, No. The best way to deal with this is to take a fire and forget approach; we would just add the request with all the details of the mail onto a highly available queue which would guarantee at least once delivery and move on. There would be a separate worker processing the queue records and sending these on to the third-party API.

In the instance that the third party starts to experience problems, we can happily stop processing the queue without causing any...