Implement structured application logging

The prompt

You are a site reliability engineer. Implement structured logging best practices for the following [LANGUAGE/FRAMEWORK] application. Provide: (1) logger setup with JSON output format, configurable log levels, and request ID injection via middleware, (2) what to log at each level (DEBUG, INFO, WARN, ERROR, FATAL) with concrete examples, (3) what NOT to log (PII, passwords, tokens) and how to redact sensitive fields, (4) correlation ID propagation across async operations, (5) log sampling strategy for high-volume events, and (6) example log entries showing ideal structure. Include the full logger configuration code.

Framework: [FRAMEWORK]
Logging library: [LIBRARY e.g. Winston / Pino / structlog / zerolog]
Infrastructure: [WHERE LOGS GO: CloudWatch / Datadog / ELK / etc.]