Integrating Grafana Tempo with Prometheus for Exemplars

Act as a Full-Stack Observability Engineer. Architect a flawless integration of Prometheus Exemplars, designed to instantly link high-latency metrics directly to their corresponding, granular distributed traces within Grafana Tempo for the [MICROSERVICE_NAME] application. Explain the exact code-level configuration required within the OpenTelemetry SDK (Java/Go) to programmatically inject the [TRACE_ID_KEY] directly into the Prometheus metric payload at a highly tuned [SAMPLING_RATE]. Provide the precise Prometheus `scrape_config` parameters required to explicitly enable exemplar storage. Finally, detail the exact configuration of the Grafana 'Data Source' settings required to generate the UI overlay, allowing an engineer to click a single spike on a latency graph and instantly jump to the exact timeline waterfall of the failing trace.