<p>Microservices-based applications are widely adopted for their scalability and support for continuous integration. While their modularity reduces development complexity, it increases operational and maintenance challenges, leading to more frequent and intricate system malfunctions. As cloud-native applications (CNAs) become increasingly sophisticated, traditional monitoring solutions prove inadequate, raising the risk of failures. These failures can include performance degradation, service outages, scalability limitations, and even potential data integrity. This has led to the concept of observability, defined as the ability to understand and diagnose a system’s internal behavior by analyzing its external states. In CNAs, resource allocation is dynamic and dependent on factors like system health, making observability crucial not only for monitoring application health but also for triggering auto-scaling when necessary. Although some studies explore observability concepts, tools, and challenges, there is no comprehensive taxonomy. This paper proposes a taxonomy for observability in microservices-based applications, with a specific focus on auto-scaling, supported by a systematic mapping study (SMS). We analyze 66 studies published between 2019 and 2024, providing a thorough overview of observability. Additionally, the SMS highlights tools, benchmarking applications, real-world datasets, and ongoing issues in the field. This work aims to offer researchers an up-to-date and comprehensive resource for understanding observability in the context of CNAs.</p>