use-red-methods

SLO design requires selecting achievable metrics based on service request rates, error rates, and durations. The RED methodology guides this process by focusing on these specific request metrics to define and set measurable performance targets.

The USE methodology focuses on resource metrics like CPU, memory, and I/O utilization, saturation, and errors. RED targets request metrics including rate, errors, and duration. Combining both provides comprehensive coverage for diagnosing performance issues.

Yes, applying the RED methodology requires access to observability tools that provide visibility into request metrics such as rate, error rate, and duration. These tools supply the data needed to diagnose performance degradation and latency spikes.

Start diagnosing slow services by checking relevant resources for saturation first, then for errors, and finally for utilization. This structured approach identifies whether resource limits or error conditions are causing the performance bottleneck.

Apply these performance investigation methodologies during capacity planning to evaluate resource utilization and saturation limits. Identifying error conditions and bottlenecks early helps predict system limits and prevent future performance degradation.