A structured log is a log entry written in a consistent machine-readable format, such as JSON with named fields, instead of free-form text. Rather than one sentence describing an event, the entry has fields like timestamp, severity, service name, request identifier and message, which tools can filter and aggregate reliably.

The benefit is searchability. Questions such as which requests for a particular customer failed in the last hour can be answered by querying fields directly, and entries can be linked to traces through a shared identifier, instead of relying on fragile text matching.

Structure does not fix every problem. Logs can still be too verbose or too expensive, and fields with an unbounded number of distinct values can strain some back ends. Inconsistent field names across teams reduce the benefit, and sensitive data written into fields, such as personal details or credentials, creates the same risk as in plain text logs.

To use them well, agree a common set of field names, include a request or trace identifier, keep sensitive values out or redact them, choose log levels deliberately and review which logs are actually queried so that unused volume can be reduced.