A cloud engineer must find the root cause of a recent performance issue.

In Cloud+ (CV0-004) troubleshooting and observability objectives, logs and metrics provide valuable signals, but they often don’t show where time is spent across a distributed request path. When an application spans microservices, APIs, queues, and databases, metrics can indicate symptoms (high latency, CPU, error rate) and logs can show events, yet correlating them end-to-end can still leave uncertainty about the true bottleneck. Tracing adds the missing “third pillar” of observability by following a single transaction through each component and measuring per-hop latency, dependencies, and call sequences. This makes it the best method for pinpointing the source of performance degradation (for example, a slow database call, an overloaded downstream service, or excessive retries). Aggregating logs is helpful for central visibility, but it doesn’t inherently provide request-level timing across services. Verbose logging increases overhead and noise and is typically used after narrowing scope. Code review is slower and not the first step when operational telemetry can directly isolate the bottleneck.