Go
In this lesson, we'll be starting our discussion about Go and its relation to microservices. Let's begin!
Microservices and the increasing popularity for Go#
Go is a programming language that is increasingly being used for microservices due to its great speed and support for concurrency. Concurrency enhances the efficiency of using multiple machines and cores.
Go also provides a powerful standard library for the creation of web services.
For further details on how Go compares with the other 4 languages commonly used for implementing Microservices, visit this site.
Similar to Java, Go is based on the programming language C. However, in many areas Go is fundamentally different from C.
Go code#
The Go program below responds to HTTP requests with HTML code.
Line 3:
The key word import imports some libraries, among others for HTTP.
Line 10, and 23:
The program’s main function defines which methods should respond to which URLs. For example, the method Footer (line 23) returns HTML code.
On the other hand, for the URL /common/css (line 10) the application delivers content from files.
It is also very easy to implement a REST service with Go.
In addition, libraries like Go kit offer many more functionalities to implement microservices.
Go build and compilation#
Go compilers are particularly well suited for Docker environments because they can create static binaries.
Static binaries do not require any further dependencies or a specific Linux distribution.
However, the applications must be compiled to Linux binaries. This requires a Go environment that can create Linux binaries.
Docker multi-stage builds#
The example uses Docker multi stage builds. Such a build divides the build process of the Docker image into several stages.
First Stage
The first stage can compile the program in a Docker container with a Go build environment.
Second Stage
The second stage can execute the Go program in a Docker container as a runtime environment that contains only the compiled program.
Consequently, the runtime environment has no build tools and is therefore much smaller.
Docker multi stage builds are not very complicated, as a look at the Dockerfile shows:
Stage 0#
Line 1
The base image golang contains the Go installation.
Line 3
The Go source code is copied and compiled into this image (line 4/5). With that, stage 0 of the build is finished.
Stage 1#
Stage 1 creates a new Docker image.
Line 7
The image, scratch, is an empty Docker image.
Line 8 and 9
The Dockerfile copies the bootstrap library (line 8) and the compiled Go binary from stage 0 (line 9) into this image.
The option --from=0 indicates that the file common originates from stage 0 of the Docker build.
Stage 2#
Line 10
Finally, ENTRYPOINT defines the binary that is supposed to be started.
Line 11
CMD indicates that no options are to be passed to the binary at the start.
Normally, ENTRYPOINT would be a shell that starts the process that is configured with CMD.
However, in the scratch image, there is no shell.
Line 12
According to Docker Documentation, “The EXPOSE instruction informs Docker that the container listens on the specified network ports at runtime. You can specify whether the port listens on TCP or UDP. The default is TCP if the protocol is not specified.” So, port 8180 is specified here.
Q U I Z
What are the advantages of using multistage builds to compile Go code?
A)
They are much faster
B)
They result in runtime environments that are smaller and more secure
C)
They are not only the industry standard for Go, but Go code cannot be compiled any other way
In the next lesson, we’ll discuss the potential of Go Lang in the implementation of microservices.
