Network slicing is a concept to provide dedicated and isolated telco services tailored to specific needs of different verticals over a common infrastructure. Offering service assurance (SA) in network slicing implies full awareness and quick response to all frequent changes in topology, resources, function operation, service requests, and slice performance. It requires a precise coordination between all components of network slicing to facilitate efficient access, processing, and analysis of the information necessary to take right actions in an architecture that is new. To address this open issue, we study those new architectural components and propose a framework to realize network slicing SA. The framework joins the two key components of SA, monitoring and analytics, in a hierarchical and distributed way. These components coordinate at various levels to optimize the tradeoff between monitoring cost and analytics accuracy. Detailed workflows are presented to demonstrate how the framework is implemented in a modular way across different layers and domains. This framework can be used to guide the integration of diverse SA solutions into a unified SA solution.