When establishing Service Level Objectives (SLOs) for a web application, what should the proposed latency SLOs be based on current customer satisfaction?

The proposed latency SLOs should reflect user experience data, particularly how latency impacts customer satisfaction. A well-crafted SLO aims to balance performance expectations with what customers have indicated is acceptable based on current satisfaction levels.

The choice indicating 90 percentile at 150ms and 95 percentile at 300ms is appropriate because it suggests a level of performance that aligns with industry standards for web applications while being realistic based on customer feedback. Typically, for a web application, users expect fast responses, and maintaining latency that ensures a majority of users (at the 90th percentile) experience good performance is critical.

Setting the 90 percentile latency at 150ms accommodates expectations for responsiveness while allowing some flexibility in terms of peak demand handling, and the 95 percentile at 300ms still keeps a relatively positive user experience.

The other options either suggest too aggressive latency targets or set benchmarks that could lead to user dissatisfaction. If the thresholds become too stringent (for instance, under 120ms at the 90 percentile), it may not be achievable in practice, especially during high load times. Conversely, options with higher thresholds (like 250ms and above) for the 90 percentile may risk exceeding what many users find acceptable, leading to disengagement or frustration.