What SLI should you choose to measure the user experience for a microservice-dependent application?

Choosing a quality SLI that measures the ratio of non-degraded responses to total responses is essential for gauging the user experience in a microservice-dependent application. This type of SLI directly correlates with user satisfaction by focusing on the effectiveness of the application from the users' perspective.

Microservices often interact with one another, and a single degraded service can significantly affect the end user's experience. Therefore, assessing how many responses are fully functional versus those that are degraded provides clear insight into the overall user experience. This metric helps in identifying issues in specific microservices that might be causing a degradation in service, enabling teams to pinpoint performance bottlenecks and improve the reliability of the application.

In contrast, measuring total microservices availability or response times focuses on infrastructure or performance metrics that do not necessarily reflect real user interactions or satisfaction comprehensively. While those metrics are important for operational health, they lack the user-centric view provided by the quality SLI. A capacity SLI, which measures maximum load, is more aligned with system capability rather than the quality of the user experience. Thus, focusing on the ratio of non-degraded responses gives a more accurate representation of how users perceive the application, making it the most appropriate choice for this scenario.