Modernizing Test Automation for Exelerate's Falcon Traffic Platform

DBB Software's QA engineer joined Exelerate's in-house team and rebuilt the test automation framework that runs the platform's release cycle. The modernized framework, built on Playwright and TypeScript, dropped the full regression cycle from 5 business days to 5 hours.

Industry

Technology

Service

Testing and QA

Team

1 QA Engineer

Project State

April 2023 - Ongoing

Country

Israel

About the Client

Exelerate Smart Traffic Ltd. is an Israeli company that develops Falcon, a SaaS platform for traffic-light management and control. Falcon's family of AI-based algorithms handles time-of-day plans alongside semi-adaptive and adaptive control scenarios for smart cities, with integrations into city-level air-pollution monitoring systems.

wm-strategy.com

The Challenge

Exelerate brought in a dedicated QA engineer from DBB Software to take over the existing test automation work on a flagship platform. The existing automation project had become hard to maintain due to outdated packages and conventions, and the team needed someone to bring it up to date and expand its coverage as the platform evolved.

Run and Extend the Existing Test Automation

Take over the E2E tests on Falcon, run them on each cycle, and add coverage as the team ships new features.

Modernize the Test Automation Framework

Replace the outdated packages and conventions in the previous automation project with a current framework that the team can maintain over time.

Cover Critical Platform Modules

Build test coverage for the parts of Falcon where regressions hurt the most: Junction Management, the Traffic Scheduler, Real-time Alerts, and the Decision Support System (DSS) algorithms.

Support Manual Functional Testing

Validate new features by hand, prepare test plans alongside the developers and business analyst, and run pre-release manual regression cycles so the team can hit tight release deadlines.

Solutions We Delivered

The scope grew from running the existing tests into a full QA workstream that the team relies on through every release cycle.

Modernized Test Automation Framework on Playwright and TypeScript

Rebuilt the legacy automation project on a current toolchain: Playwright as the test runner, TypeScript for typed test code, API-driven test data generation and state sharing, and Gitlab CI integration for scheduled rules. The new framework replaced the outdated packages and conventions in the previous setup with a structure that the team can extend without fighting the framework on every change.

Coverage Across Critical Falcon Modules

Built and maintained 60+ automated E2E and API tests covering the platform modules where regressions hurt the most: Junction Management, the Traffic Scheduler, the Real-time Alerts subsystem, and the Decision Support System (DSS) algorithms. The Playwright-based API layer generates test data on demand, so complex scenarios run end-to-end without setup through UI between cases.

Dockerized Nightly Runs Through GitLab CI

Set up automated test execution in Docker containers, triggered through GitLab CI as Nightly Runs across the staging environments. The team gets fresh regression results each morning instead of waiting for the next manual cycle, and 100% of the critical functions stay under automated check between releases.

Manual Testing and Pre-Release Regression

Validated new features by hand, prepared 20+ manual test plans alongside Exelerate's developers and business analyst, and ran the full pre-release manual regression cycles that Exelerate keeps as a release-gate policy. The manual workstream sits alongside the automation rather than replacing it, which is how the team wanted the QA setup to work.

Results Achieved

Full Regression Cycle Dropped From 5 Business Days to 5 Hours

The new automation framework handles the regression cycle end-to-end on staging, taking what used to be a five-business-day manual cycle down to roughly five hours of automated runs.

Critical-Flow Automation Dropped From 12 Hours to 2 Hours

Specific to the critical-flow tests carried over from the legacy automation project: runtime dropped from around 12 hours on the previous setup to about 2 hours after the framework rebuild.

100% Automation Coverage Across Staging

Critical functions across the staging environments sit under automated checks, with Nightly Runs in Docker through GitLab CI catching regressions overnight.