An SRE team is building an observability strategy for a new microservices platform. They want to (1) alert when a service's error rate crosses a threshold over time, (2) inspect the exact stack trace and message emitted when a specific request failed, and (3) understand how a single user request flowed across six services to find where time was spent. Which combination of telemetry signals correctly maps to these three needs, in order?