Effective Logging Practices for Java Backend Services

The author revisits previous articles on poor Java code and interface/controller standards, then highlights a frequent issue: production bugs that cannot be reproduced due to insufficient logging, especially in large clusters where locating the responsible node and user actions is painful.

Two essential logging requirements are proposed: (1) the ability to identify the specific machine handling a request, and (2) the ability to capture what the user did. To satisfy the first requirement, the Nginx configuration is adjusted to return the processing node in response headers, with a simple example configuration shown.

For the second requirement, the author recommends using Log4j's MDC (Mapped Diagnostic Context) to store user information per request, clearing it after the filter to avoid thread‑pool contamination. Sample filter code and MDC usage are illustrated.

After establishing these foundations, three concise logging rules are defined for developers: (1) all modifications (including additions) must be logged, ensuring data changes are traceable; (2) conditional branches must log condition values and important parameters, exemplified by printing a userType value that determines code flow; (3) when processing large data volumes, log the amount of data processed to aid performance analysis.

The article also presents example log outputs and emphasizes that log levels are less critical than consistent logging practices. It concludes with advice for newcomers: rely on logs rather than debugger, review logs before committing code, and treat logging as an integral part of functional development.