Correlation IDs Across Services
What you'll learn: How to connect logs from multiple services that collaborate on a single user request using trace and span identifiers.
The Challenge of Distributed Logging
When a user request flows through five microservices, each service writes its own logs independently. Without a linking mechanism, you'd need to manually search each service's logs by timestamp and guess which entries belong together—a nightmare during incident investigation.
Trace IDs and Span IDs
A trace ID uniquely identifies the entire user request journey across all services. Think of it like a package tracking number that follows your shipment through every warehouse and truck.
A span ID uniquely identifies each step or service call within that journey. Each service generates its own span ID but inherits the parent's trace ID.
When Service A calls Service B:
- Service A generates a trace ID (if starting the request)
- Service A creates a span ID for its work
- Service A passes both IDs to Service B via HTTP headers
- Service B logs both IDs and creates its own child span ID
- Now you can query logs for that trace ID across all services
In Practice
Every log entry includes fields like:
{"trace_id": "abc123", "span_id": "span-001", "message": "Processing order"}
When investigating, you search all services for trace_id: abc123 and instantly see the complete request path in chronological order.
Key Takeaway: Propagating trace_id and span_id fields through service calls transforms isolated log entries into a connected story, enabling rapid debugging of distributed system failures.