Continuous Testing in the World of Agile and DevOps

Yuri Filatov
7 min readJul 8, 2022

How continuous testing improves Agile and DevOps

As businesses embrace Agile and DevOps, continuous integration (CI), delivery (CD), and testing (CT) are becoming key elements that help to improve the product. According to the latest research, organizations using this pipeline in a software development life cycle deploy code 208 times more frequently, fix bugs 2.6 times faster, and get 7 times fewer refusals from users. All this increases their credibility in the market where speed and quality have become the main advantages. Let’s figure out the role of continuous testing in companies operating according to Agile and DevOps principles and why it should not be ignored.

Continuous testing in CI/CD

Both Agile and DevOps are intended to deliver high-quality software on time. Agile focuses on optimizing the development life cycle, while DevOps integrates programming and operations in a CI/CD environment. When applied together, both methods optimize software development and result in creating better products.

Continuous integration, testing, delivery, and deployment are part of the CI/CD concept corresponding to an automated pipeline. This pipeline facilitates the process of merging newly-written finished code with the main code base, runs various tests and checks, and starts automated deployment. All this depends on how deeply CI/CD is integrated into a project.

The structure of CI/CD resembles a matryoshka doll, where continuous integration is complemented and extended by testing, delivery, and deployment. The aim of these processes is to improve the development reliability and stability of releases while reducing terms.

Let’s take a look at an entire development cycle that includes continuous testing.

The cycle begins with writing code and covering it with tests. Then, a system like Jenkins automatically collects changes and starts testing them. After automated tests are successfully passed, the changes go to manual testing. The result is a stable version of the product assigned with a number, and now these changes are called a release candidate.

Next, the release branch is uploaded and deployed on the client’s production server. The next-to-last step is the conduction of monitoring, which is when engineers keep an eye on the deployed version of the product, stabilizing or fixing the application if necessary. The cycle ends with planning new functionality or collecting data on changes and adjustments for future releases.

The CI process includes the aforementioned automation of steps for verifying the newly written code. Development is conducted in separate branches, and then all changes are merged into the main branch. The process of assembling and testing starts automatically and takes as little time as possible.

The idea behind the CD is to automatically perform the continuous integration step and prepare the project for release to the production server. Usually, a DevOps specialist joins the development process at this stage. They take a build that has been tested and proven at the CI stage and deploy it.

What continuous testing includes

Continuous testing is a small cog in the cycle of continuous development, integration, and deployment. It presupposes the process of performing automated tests at all stages of the software development life cycle — from planning to production. This allows a business to receive feedback on risks as quickly as possible. Simply put, continuous testing answers the question “what is the quality of our software right now?”

Software testing services of the old type proposed a phased handover of software from one team to another. Plus, a project used to have certain phases for development and quality control.

But the traditional way of getting feedback manually at each development stage leads to longer integration cycles and delayed product release. That’s why a need for continuously testing software appeared.

In today’s reality, testing can’t be an isolated process, excluded from the application release. Teams must be able to work together effectively and efficiently, and quality becomes a shared concern of developers, QA engineers, and operations teams.

In a DevOps-based development process, code is developed, delivered, tested, and deployed continuously. For example, every time a developer checks code in a repository, an automated suite of unit tests is executed in a continuous process. If a build passes the checks, it is deployed to QA performance servers for comprehensive functional and load tests. When the tests are completed, the application is launched into production.

The aim of continuous testing is to assess the quality of software at every stage of the continuous delivery process through early and frequent testing.

Why is the old model no longer effective? Application architecture is becoming more distributed and complex, spanning the cloud, APIs, microservices, and so on. This creates almost endless combinations of different protocols and technologies within a single business transaction. At the same time, many programs require very frequent releases — up to a thousand times a day, as happens at Amazon, Google, and Netflix. That’s why the time for developing tests is drastically reduced.

In addition, a defect in an app operation often leads to business problems. This, for example, happened with HSBC — they had a major software failure in January 2016, and millions of clients couldn’t access their online accounts. The bank managed to recover and return to normal functioning in almost two days, but its credibility was undermined.

Testing used to be considered the enemy of high speed and was of questionable business value. But today, by using continuous testing, any software testing company can help to establish smarter processes focused on frequent releases, quality, and security of software.

What the role of test automation is

Automated testing and continuous testing complement each other, but they are not to be equated. Each of these processes has a huge impact on DevOps and continuous delivery in its own way.

In test automation, a dedicated tool to complete a set of tasks faster and with fewer errors is used. Typically, after a feature is developed and integrated into a centralized repository, a tester starts writing scripts, which significantly slows down software delivery.

Continuous testing is more extensive than automated testing. As book author and software researcher Jez Humble thinks, its goal “is to run many different types of tests — both manual and automated — continually throughout the delivery process.” Here, test scripts are written before the function is developed. As soon as a programmer integrates the source code of the new functionality into the centralized repository, various test scripts assessing the code quality are automatically run.

At its core, continuous testing is a methodology that uses test automation to improve efficiency and accelerate software delivery.

Benefits of continuous testing

Incorporating continuous testing into DevOps processes gives companies the following benefits:

  • Prompt feedback

Continuous testing automates quality assurance and workflows at every stage of the software development life cycle. Feedback gives specialists the information needed to improve the code before it is deployed.

Researchers from the National Institute of Standard Technology have found that fixing a bug detected during the software production stage takes 15 hours. This is three times longer than if the same bug was detected at the coding stage.

A famous example of inaccurate testing was the failure of the Mars Climate Orbiter. Due to defects in software, the orbiter collapsed as soon as it entered the atmosphere of Mars. This fiasco cost NASA approximately $125 million.

Early feedback helps to promptly assess potential business risks — from a negative change in user experience to unsuccessful software releases.

  • Fast delivery

As we said above, all processes occurring after the code is integrated into the repository are planned and automated even before the development is completed. In a non-DevOps practice, testers spend a lot of time (50%) finding suitable test data, and 70% of tests are still developed manually. These figures can be significantly improved by organizing software testing outsourcing with the use of automation platforms.

  • Improved quality control

Thanks to continuous testing, teams can deliver quality assurance services at a higher level. The use of automated control tools ensures that functions, having failed validation at one stage, won’t enter the next. For example, Google runs about 150 million tests daily, and almost all of them are automated. Boris Chen, Co-Founder and Vice President of Engineering at tCell, claims that automation not only saves one person’s time but also allows the entire team to work faster.

Source: andersenlab.com

  • Minimized risks for a business

An error in even one unit of an app can result in unnecessary business downtime and negatively impact productivity and profitability. This is especially true for companies that depend on the availability of services.

Continuous testing reveals issues that could be missed in large software systems, and reveals them in full detail. This helps to avoid the costs associated with business disruption.

Source: andersenlab.com

Why continuous testing is important to DevOps and Agile

DevOps is all about ensuring continuity — whether it’s an integration of code, launch of tests, or deployment across multiple environments. In this chain, continuous testing is a bridge that unites continuous integration with continuous delivery and deployment to create an efficient software development pipeline.

This workflow ensures that the new feature will meet the expected quality criteria and that the application can be given to the user on time. Continuous testing is aimed at reducing the amount spent on low-quality software, which accounted for $2.08 trillion in 2020 in the US alone.

The results of recent Forrester Total Economic Impact™ (TEI) research illustrate how several corporate companies have earned over $19 million by introducing the practice of continuous testing. Enterprises have cut test case development time by up to 50%, increased software release productivity by 40%, and accelerated deployment of critical code from 8 hours to 30 minutes.

Continuous testing makes Agile and DevOps methodologies more effective and efficient in ensuring fast software development. It guarantees profitable and speedy releases by finding defects early in the development cycle, improving the process, and reducing project costs. And this, in turn, means that software delivery becomes more flexible, high-quality, and rapid.

--

--