Postbot increases API test coverage by 24% in the first month of adoption for BAM Technologies

Digital transformation presented BAM with the initial challenge. In 2021, BAM moved from a three-layered architecture consisting of a backend, frontend and database to a microservices structure.

Previously, there was no centralized location for requests. The old solution generated Swagger, which was convenient for calling an API with one or two requests in the body. This process required the developer to copy and paste, pull data from elsewhere and manipulate an authorization token. It was cumbersome, time-consuming, and prone to human error. The goal was to make the process scalable and efficient.

This new infrastructure required comprehensive, automated end-to-end (E2E) testing using a process that was easy to scale. Additionally, documentation needed to be easy to access and manage.

BAM started using the Postman API Platform to provide API endpoint documentation, consuming Swagger output from its dotnet microservices.

“The Postman API Platform has all the essential functions a developer needs to efficiently and effectively work with APIs.”

Ewin Hong, Platform Quality Control Lead, BAM Technologies

Management of the microservices is now seamless, with Postman providing automation, consistent documentation, comprehensive testing and easy access to requests. Uncertainty and second-guessing have been eliminated.

API documentation has been considerably enhanced. It's a single source of truth, and systems architects can now better implement API structure best practices. All developers can follow the same documentation, leading to greater consistency in API structure across the business. Consequently, API governance has also improved.

“Postman is the equivalent of a cookie cutter for APIs, helping to bake APIs that are structurally sound and don’t crumble.”

Ewin Hong, Platform Quality Control Lead, BAM Technologies

Documentation has increased developer productivity when building APIs. Developers, especially those from different teams, no longer spend time trying to work out the API structure.

Mock servers are essential to the API design process. They are used to design the initial endpoint request structure, including the URL, query/path parameters, request body, and expected output. A senior developer then reviews the API and corrects anomalies. The API is then approved and built.

BAM also uses a team workspace to store collections for microservices and E2E testing. The E2E test collection includes a health check, microservices call authorization, and a preliminary set-up collection, with a pre-request script for calling an authentication service, that can be used for all microservices.

There are several teams, so each subteam forks a collection from the main workspace to its own team's separate workspace. In this subteam workspace, contributions and updates are made. Once work is complete, the subteam generates a pull request to the main E2E collection, which is reviewed and approved by a senior developer. The team workspace provides visibility, improves efficiency and enhances API collaboration between teams.

Since adopting Postbot, BAM has used the AI assistant to drastically increase test coverage with collections and requests. Firstly, logical tests are generated, and are then adapted to meet business requirements. With Postbot, developers are given inline autocomplete suggestions while manually editing code. It also generates API documentation with a single click.

The first microservice built in 2021 was an authentication microservice. By early 2022, the second microservice was completed, with Postman being implemented midway through the year. The creation of microservices accelerated, with 30 microservices in operation by June 2023.

Postbot has contributed to enhancing BAM's product reliability by improving testing. This has led to higher-quality code and structurally sound endpoints, meaning microservices function as expected. Since using Postbot, BAM has been able to double the number of microservices with scalable tests.

One month after adopting Postbot, API test coverage jumped from 17% to 41%, an overall total increase of 24%. This almost quarter increase in test coverage clearly demonstrates that more requests are being sent, and consequently, more tests are being created for those requests.

Collections, workspaces, and mock servers have improved developer productivity and collaboration. For example, using mock servers means all the APIs have a set structure, enabling developers to free up time to focus on innovation.

Another clear indication of operations scaling has been the dramatic increase in API document creation, which has doubled. Now, more team members than ever before are reading those documents, encouraging collaboration and innovation.