The client – a worldwide vendor of accurate data for streamlining a decision-making process and ensuring a deep understanding of the automotive industry – has occupied a solid niche on the market. With offices in 40+ countries, the customer delivers solutions based on research datasets for automotive manufacturers globally.
More than 35 years of data gathering have encouraged the company to build a new platform for its storage and migrate all the information to the cloud in order to ensure further back-end for new products. This solution encompasses non-relational databases, message queues, their processing by Azure Functions, and many more.
Therefore, the client required thorough testing to ascertain the quality of the delivered product.
In order to help the client meet the core business objective, a1qa assigned 7 FTEs. The QA team smoothly integrated into the established Agile delivery ecosystem.
To be precise, SAFe framework was in use because of a geographically dispersed product delivery team (the UK, the USA). As the platform consists of different dependent subsections, the teams rely on each other while delivering new functionality. Due to SAFe implementation, the overall performance was well-coordinated, and possible risks were mitigated already on the planning stage of each PI that lasted for a month (previously for 3).
a1qa was responsible for providing functional testing and test automation services.
Prior to testing itself, the engineers were assessing user stories to spot any logical contradictions, fix them at once, and approve. A defect was opened only in case of making it in production. Such a careful approach resulted in a small number of occurred issues.
In terms of functional testing, the engineers were verifying the correct operation of API in terms of multiple components of Microsoft Azure Cloud.
Within the non-relational database, all the information is stored in documents presented in a JSON format. The a1qa engineers were looking for a particular data type in the original database, verifying data connections, and checking for compliance with the set model.
Test automation was applied regarding the whole regression testing scope on the project so there were only automated acceptance tests.
A distinctive feature in the process concerned the fact that QA automation engineers were closely interacting with developers and implementing tests already during a sprint. User stories were completed only after automated tests were created.
Initially chosen test automation framework wasn’t suitable for meeting the set QA needs, therefore, a1qa in cooperation with developers created a new solution based on SpecFlow 3. However, the client asked to shift to test implementation applying NUnit. It didn’t require Gherkin scenarios where each separate step complied with a particular part of code. So the code prepared by the a1qa engineers was similar to that written by developers.
Cooperation with the client continues. Currently, a1qa is introducing solutions to streamline the testing process, e.g. MockServer for ensuring the quality of microservices in isolation and their interaction.