Have you been tasked with implementing test automation across multiple systems? Are there already islands of automation, but different teams have different solutions, frameworks, and tools? Don’t worry, we’ve been in the same situation many times, and we have a strategy to handle it! With our experience and our lightweight framework Tzatziki, we bring everything together and bring order to reporting and execution.

Through our strategy of reuse, authentication, and test data management, we allow teams to focus on what they do best. The portal team, responsible for coding the end-user website, has full knowledge of the interface, and their regression tests can be easily adapted to fit into a longer flow that includes billing and delivery checks. The purchased billing system offers ample integration APIs, allowing us to write our own automation code there. The delivery system is outsourced and requires real addresses to function. In that case, we need to book test data before each test run. No problem!

In workflows spanning multiple systems, managing numerous identities can quickly become complex. When expressing everything as code, there is a temptation to embed passwords directly in the code. However, this approach is incorrect. In Tzatziki, we use a password vault for this purpose. HashiCorp Vault provides an easy-to-operate and scalable solution for managing passwords and other shared secrets, making it equally convenient for use during test automation development and execution within CI/CD pipelines. In Tzatziki, we securely and seamlessly handle authentication against Vault to minimize disruptions.

Teams developing test automation often have varying preferences when it comes to tool selection. Many times, the key to success lies in allowing teams to choose their own tools. However, a common trait of good tools is their support for CI/CD and their ability to run with different parameters in Jenkins Pipelines, GitHub Actions, Azure DevOps Pipelines, or similar platforms. With minor customizations and with the help of Tzatziki, these jobs can become steps in a longer flow, and the right teams with the right skills maintain the appropriate parts of the tests.

In more complex flows in integrated environments, test data quickly becomes a bottleneck. Data often needs to be kept synchronized across multiple systems for the whole to function as intended, and the risk of encroaching on colleagues’ test data increases when tests are performed automatically. Tzatziki’s strategy is to maintain a simple internal database for test data where data can be booked from external systems, enriched before tests, reserved for various test cases, and then cleaned up and returned after use.

Java is a requirement, but beyond that, Tzatziki’s support isn’t reliant on the test execution engine. However, we’ve found Cucumber to be exceptionally expressive after decades of development. By choosing Cucumber not only for unit tests but also for requirements gathering, reporting, and handling longer, more complex test cases across multiple systems, we achieve a broader understanding among stakeholders within the company. Cucumber is well-supported in most popular programming languages and testing frameworks, which facilitates both execution and reporting. Tzatziki is well-suited to complement and streamline test automation with Cucumber for Java. Whether tests are run locally by developers or on the build server, the process remains straightforward. When tests are conducted at the behest of a testing organization rather than as part of an integrated build process, Tzatziki also helps avoid redundant test execution based on previous status, simplifying test suite scheduling.

Jira, which has become the de facto industry standard, is a great tool for managing development teams, handling requirements and follow-ups, and compiling test reports with various add-ons. Choosing a test management add-on often comes down to personal preference. We are particularly fond of XRay for its balanced mix of flexibility and functionality. XRay naturally supports tests expressed in Cucumber format. With Tzatziki, it’s easy to report results to multiple test plans based on the tagging of test cases. When Jira’s dashboards aren’t sufficient, Tzatziki has support libraries for reporting directly to Confluence.

Tzatziki is intended to be the sauce that complements the work of test automation. Given that the team already understands the basic principles behind tools like Cucumber for Java and is comfortable with OpenAPI, Selenium, Playwright, and other frameworks, Tzatziki takes care of what’s missing. Tzatziki is not a framework that completely takes over and dictates how tests should be written and implemented, but rather allows the team to focus on what really matters: test automation.

At Lemontree, we have worked with the principles behind Tzatziki for many years in various forms. We have experience in setting up and maintaining the infrastructure required. We know which strategies facilitate maintenance over time, and we are pragmatic problem solvers with competencies in both product development and testing. Tzatziki is a dynamic framework that continuously expands with support for new components as we encounter technologies that benefit many. Experience from one project benefits others. In addition to the framework code, we have a detailed strategy for modularizing the test code, dividing test cases for test data enrichment, and reusing them across multiple stages.

johan.klintholm@lemontree.se | LinkedIn