3 minutes read

POSTED Apr, 2021 dot IN Debugging

Microservices Testing: A Quick Start Guide

Burak Kantarcı

Written by Burak Kantarcı


Product Manager @Thundra

linkedin-share
 X

What is Microservices Testing?

A microservices architecture creates an application as a collection of services. Each microservice works independently and uses application programming interfaces (APIs) to communicate with other services. Each microservice has its own data store and is deployed independently.

Testing a microservices application requires a strategy that accounts not only for the isolated nature of microservices but also for service dependencies. This is why microservices testing processes often put each microservice in isolation to verify it is working properly and then tests microservices together.

Microservices testing is becoming a central part of continuous integration / continuous delivery (CI/CD) pipelines managed by modern DevOps teams.

In this article, you will learn:

  • Microservices Testing Types
    • Unit Testing
    • Contract Testing
    • Integration Testing
    • End-To-End Testing
  • Microservices Testing Strategy
    • Pick the High-Risk Services
    • Use Service Virtualization
    • Test to a Service Level Agreement
  • Microservices Testing with Thundra

Microservices Testing Types

When developing a microservices architecture, there are several tests you can run to validate your application, including unit testing, contract testing, integration testing, end-to-end testing, and UI/functional testing.

Unit Testing

This type of testing helps you validate the performance of each component (or unit) of your software. A unit can be an object, a module, a procedure, function, or method. You can define the size of the unit but be advised that smaller units are the easiest to test. Furthermore, if you are struggling to define a unit test, consider it an indication your module should be broken down into smaller pieces.

Ideally, developers should perform unit testing during the development phase, to ensure all code components are working properly and as designed. However, during rushed development or when teams are working in siloed phases, the quality assurance (QA) team may be required to perform unit testing—after the developers completed their work on the project.

Contract Testing

To ensure that services can communicate, a contract defines the results or outputs expected for specific inputs. Consumer-contract testing ensures that each service call fulfills that contract, and that, even if the service itself changes internally, consumers will continue to receive the same results from that service.

The service itself thus remains a black box. To test the contract, services are called independently and responses are verified. If there are service dependencies, they must operate as stubs that enable a service to function, but without interacting with other services. This will also prevent erratic behavior caused by external calls, putting the focus on testing a single service.

Integration Testing

An integration test collects microservices and tests them together as a subsystem. The goal is to verify that the microservices collaborate to achieve a joint operation. During the test, communication paths are exercised, making it possible to locate incorrect assumptions in each microservice, regarding interactions with its peers.

Integration testing can also help you verify the interactions between microservices and external components. For example, you can use this type of test to verify that external data stores and caches are properly integrated, respond, and perform as expected.

End-To-End Testing

The main goal of an end-to-end test is to verify that the entire application works to achieve its business objectives, regardless of the components making up the entire architecture. During end-to-end testing, the application is treated as a black box while tests exercise the system as it is fully deployed.

End-to-end testing is especially important for microservices architecture, which is composed of many parts that attempt to achieve similar behaviors. An end-to-end test can help you ensure there are no gaps between microservices and all work towards achieving their goal. Ideally, the tests should verify messages pass correctly between services, and that network infrastructure, like firewalls and proxies, is properly configured.


GET YOUR FREE ACCOUNT TODAY

Unlock visibility into the CI pipeline by spotting test failures in no time.

Sign Up

3 Microservices Testing Strategies

Here are three strategies you can use to make microservices testing more effective.

Pick the High-Risk Services

Continuous integration and continuous delivery (CI/CD) processes require automated testing. For microservices that must operate at a large scale, load testing is especially important, as are other types of testing that can be complex and time-consuming, such as integration testing and microservices security testing. However, these types of testing can take time and slow down the CI/CD pipeline.

The solution is to identify high-risk microservices and test them extensively at the beginning of the cycle to reduce risks later on. This may require additional time, but it will be worthwhile by reducing issues later in the process. For low-risk microservices, heavier tests can be delayed until later in the process.

Use Service Virtualization

Dependencies can create bottlenecks if the code is ready but held back because a dependency is pending delivery. You can use service virtualization to avoid this delay.  

Service virtualization lets you mock out a service. First, you need to create a microservice containing the required URLs. Next, the URLs can accept data and respond according to the microservice specification. Note that response can be limited in scope.

Here is an example of an endpoint mockup response for /customers/{id}:

{
"id": 567,
"first_name": "Jane",
"last_name": "Smith"
}

While the behavior is limited, the mockup does provide the consumable functionality of a dependency still under development.

Test to a Service Level Agreement

To ensure efficiency and collaboration throughout the development process, testers, developers, and relevant stakeholders need to be on the same page. You can make sure all collaborators are on the same page by creating a Service Level Agreement (SLA). Ideally, testers and developers should work together to create this document, and implement it with flexibility in mind.

Microservices Test Visibility with Thundra Sidekick

Testing microservices gets only harder as modern applications become much more distributed and asynchronous. Thundra Sidekick is designed to help developers to debug the applications on remote machines with non-intrusive debugging integrated with distributed tracing. Using these tools, developers can shed a light on the test failures running on build machines and/or during the CI phase. Sign up with Thundra Sidekick to get started with a better understanding of microservices testing.