In the previous chapter, we discussed delegating application responsibility to networked services accessible by API and intra-process communication and handled by a message queue.

This approach mimics an emerging trend of breaking large monolithic applications into smaller chunks; thus, allowing developers to leverage dissonant languages, frameworks, and designs.

We listed a few upsides and downsides of this approach; while most of the upsides dealt with keeping the development agile and lean while preventing catastrophic and cascading errors that might bring down an entire application, a large downside is the fragility of each individual component. For example, if our e-mail microservice had bad code as a part of a large application, the error would make itself known quickly because it would almost assuredly have a direct and detectable impact on another component. But by isolating processes as part of microservices, we also isolate their state and status.

This...