Structured JSON Logging

[mftee]

Replace the current plain-text tracing output with structured JSON logs. Structured logs are machine-readable and integrate with log aggregation systems like Datadog, Loki, and Elasticsearch.

Requirements:

• Replace tracing_subscriber::fmt() in main.rs with a JSON formatter using tracing-subscriber with the json feature
• Every log entry must include:
  • timestamp (ISO 8601)
  • level (INFO, WARN, ERROR, DEBUG)
  • message
  • service (always "stellar-doc-verifier")
  • target (module path)
• Add request-scoped context to logs using tracing::Span so that all logs within a request include:
  • request_id (UUID generated per request via middleware)
  • method (HTTP method)
  • path (request path)
• Create an Axum middleware layer RequestTracingLayer that creates a span per request and injects request_id
• Log level must be configurable via LOG_LEVEL env var (default: info)

Acceptance Criteria:

• Every log line is valid JSON parseable by serde_json
• Request logs include request_id field
• LOG_LEVEL=debug enables debug output
• Unit test parses a sample log line and asserts required fields are present