The company needed to complete a complex migration on a tight deadline to avoid millions of dollars in post-contract fess and fines.
Transitioning from Python to Golang for eBike App Development
Our client is a well-established technology solutions provider with a strong presence in multiple sectors. To maintain its competitive edge and drive innovation, the client continuously seeks to optimize its digital solutions.
This time, the client sought a solution for its eBike application, which faced significant performance and scalability challenges. Facing these limitations with its existing system, the client aimed to modernize its architecture by transitioning from a legacy Python-based application to a modern, microservices-based Golang solution. This transition was aimed at improving usability, reducing operational costs, and enhancing performance to support future growth and innovation.
The Challenges
Client's App Challenges
- Long loading times: The app’s slow performance created a suboptimal user experience.
- High costs: Maintaining and scaling the Python-based application was resource-intensive.
- Limited scalability: The monolithic structure restricted the app’s adaptability to evolving needs.
Development Challenges
The team faced several obstacles during the development process. First, the lack of documentation for the existing Python app required extensive reverse engineering to understand and replicate its functionality. Additionally, managing distributed data across microservices posed challenges in maintaining consistency, as each service had its own database. Achieving eventual consistency across services added complexity, requiring careful design to ensure all systems remained in sync over time.
Coordinating distributed transactions across multiple services was another hurdle, as failures in one service could disrupt the entire workflow. To prevent data anomalies, the team needed to implement mechanisms for idempotency, ensuring duplicate events did not affect the system’s integrity. Finally, robust monitoring and retry mechanisms were essential to address failures in inter-service communication, providing reliability and resilience in the new architecture.
Solutions Provided
The app was re-engineered from a monolithic Python application to a microservices-based architecture using Golang. This restructuring enhanced performance and scalability while simplifying maintenance.
Microservices Breakdown
- User Service: Handles profiles, authentication, and user preferences.
- Ride Tracking Service: Records ride data such as routes and durations.
- Navigation Service: Provides route planning and guidance.
- Fitness Service: Tracks and analyzes user fitness metrics.
- Security Service: Includes features like the ‘Lock’ function, which disables motor support to deter theft.
Inter-Service Communication
gRPC: Ensures efficient, low-latency communication between microservices. For example, when a ride is completed, the Ride Tracking Service uses gRPC to notify the Fitness Service to update metrics.
Addressing Data Challenges
The team addressed data challenges by implementing an event-driven architecture, where services published events to a message broker, allowing subscribed services to update data in real time and maintain consistency. Eventual consistency models ensured synchronization across services over time, with idempotent operations preventing disruptions from duplicate event processing.
To handle distributed transactions, the Saga pattern was employed, coordinating local transactions with compensating actions to preserve data integrity in case of failures. Additionally, idempotency mechanisms were designed to manage duplicate events, and robust monitoring tools like Prometheus, Grafana, and Kibana were integrated alongside automatic retries with exponential backoff to ensure reliability and seamless communication between services.
Example Flow: User Completes a Ride
- The Ride Tracking Service records the ride data and publishes a “RideCompleted” event.
- The Fitness Service, subscribed to this event, updates the user’s fitness metrics.
- The Navigation Service updates route recommendations based on the new data.
- The User Service updates the user’s ride history.
By employing an event-driven architecture with gRPC for inter-service communication, the eBike Connect platform can achieve scalability and maintain data consistency across its microservices.
Technology Stack
Golang, Python, Prometheus, Grafana, Kibana, Kubernetes for managing microservices, AWS.
Team Composition
Python Developer, Golang Developer, Product Owner (PO) and QA Engineer
