Ensuring software quality is one of the key challenges associated with the development of control applications for the powertrain domain. Software architecture plays a pivotal role in realizing functional and non-functional requirements for automotive embedded systems by breaking down the complexity of the application software. Software architecture is a work-product of the early stages of software development. Hence, design errors introduced at early stages will increase cost of rework of software artifacts. PERSIST is a model-based software product line approach focused on cross-project standardization of powertrain software facilitated by common design guidelines and adherence to standards like ISO 25010, AUTOSAR and ISO 26262. This paper shows a systematic approach to ensure objective, early and repeated analysis of software architecture for the PERSIST product line using metrics. The systematic approach involves definition, implementation, measurement and evaluation of the metrics for powertrain software architecture. The definition of the metrics is based on an extension of Goal-Quality- Metric approach taking automotive software development constraints into consideration. A quality model for software architecture is derived based PERSIST design guidelines, safety and quality standards. Metrics were defined to ensure coverage of various attributes such as modularity, complexity and maintainability. Additionally, some metrics applied at the implementation level can applied as early as the architecture stage depending upon the availability of relevant data. This facilitates frontloading of the verification and validation effort. The engineering effort required to implement the metrics was reduced by designing a modular toolchain structure which facilitated integration of various file formats (including third-party specifications) for architecture evaluation. The toolchain was integrated into a continuous integration framework which enabled recurrent assessment for series development projects. The toolchain was integrated into the software development workflow for various projects in the product line. One of the key problems related to objective evaluation of software architecture is to determine the thresholds for a metric based evaluation. The upper and the lower thresholds of the metrics are derived through a systematic approach. A concept for prioritizing quality goals based on various aspects such as stage of development, expected software maturity, organizational goals and project constraints is proposed. Depending upon the prioritization, weights for individual metrics are derived. The concepts for metric definition, implementation and evaluation were applied to hybrid control unit and battery management systems to determine measures to improve the software architecture. With a new approach of architecture-driven development, this paper demonstrates how early evaluation of the software architecture helps to frontload of development efforts, avoid late quality risks and hence reduce time and costs for series software development by applying systematically developed architecture design metrics combined with Continuous Integration.