Key-Value Pairs and Log Fields
What you'll learn: How to enrich log messages with structured context fields to make logs searchable, filterable, and truly useful for debugging and analysis.
The Core Idea
While structured logging gives you machine-readable format (like JSON), key-value pairs are what you actually put inside that structure. Instead of cramming everything into a single string message, you attach specific fields like user_id, request_id, error_code, or duration_ms alongside your log message.
Think of it like metadata on a photo: the image itself is your log message, but the EXIF data (location, timestamp, camera settings) provides the searchable context that makes that photo findable later.
Why Key-Value Fields Matter
When you log "User login failed", you're stuck. Which user? What error? Now compare this with fields:
{
"message": "User login failed",
"user_id": "12345",
"error_code": "INVALID_PASSWORD",
"request_id": "req-abc-789",
"ip_address": "192.168.1.100"
}
Now you can instantly filter all failures for that user, trace the entire request path using request_id, or analyze patterns by error_code. You're not parsing strings—you're querying data.
Common Context Fields
- Identifiers:
user_id,session_id,request_id,transaction_id - Errors:
error_code,error_type,stack_trace - Performance:
duration_ms,response_size,db_query_time - Environment:
service_name,hostname,version - Business:
order_id,product_id,payment_method
The key is consistency: use the same field names across your entire system so logs from different services can correlate seamlessly.
Key Takeaway: Key-value fields transform logs from unstructured text into queryable data, enabling you to filter, correlate, and analyze events across distributed systems efficiently.