In this chapter, we are going to deploy microservices. We will use different technologies in order to provide the reader with the knowledge required to choose the right tool for every job. First, we will use PM2 and its deployment capabilities to run applications in remote servers. Then, we will play around Docker, which is one of the most advanced deployment platforms, and the entire ecosystem around containers. In this chapter, we will show how to automate all the deployments as highly as possible.
Deployments are usually the ugly friend of the Software Development Life Cycle (SDLC) party. There is a missing contact point between development and systems administration that DevOps is going to solve in the next few years. The cost of fixing bugs at different stages of SDLC is shown in the following figure:
This diagram shows the cost of fixing a bug, depending on the stage of the SDLC
Fail early is one of my favorite concepts in the lean methodology. In the change management world, the cost of fixing a bug in the different stages of the software life cycle is called the cost of change curve.
Roughly, fixing a bug in production is estimated to cost 150 times the resources as compared to the costs to fix it when taking requirements.
No matter what the figure is, which depends a lot on the methodology and technology that we use, the lesson learned is that we can save a lot of time by catching bugs early.
From the continuous integration up to continuous delivery, the process should be automated as much as possible, where as much as possible means 100%. Remember, humans are imperfect and more prone to errors while carrying out manual repetitive tasks.
Continuous integration (CI) is the practice of integrating the work from different branches on daily basis (or more than once a day) and validating that the changes do not break existing features by running integration and unit tests.
CI should be automated using the same infrastructure configuration as we will be using later in pre-production and production, so if there is any flaw, it can be caught early.
Continuous delivery (CD) is a software engineering approach that aims to build small, testable, and easily deployable pieces of functionality that can be delivered seamlessly at any time.
This is what we are aiming for with the microservices. Again, we should be pushing to automate the delivery process as, if we are doing it manually, we are only looking for problems.
When talking from the microservices' perspective, automation on deployments is the key. We need to tackle the overhead of having a few dozen of services instead of a few machines, or we can find ourselves maintaining a cloud of services instead of adding value to our company.
Docker is our best ally here. With Docker, we are reducing the hassle of deploying a new software to pretty much moving a file (a container) around in different environments, as we will see later in this chapter.