Logging levels - meanings and use

All popular logging frameworks follow a similar severity convention for log levels. There are five primary levels used in production code, with two additional special‑purpose levels.

Order (high priority to low):

• FATAL · ERROR · WARN · INFO · DEBUG · TRACE

Meaning and use:

1. DEBUG for fine‑grained diagnostic messages that help you trace flow and variable values. These should never appear in production – set your root logger to INFO or higher to silence them. In your code: logger.debug("Processing request id={}", id);
2. INFO for normal, informative messages that mark significant application events (startup, shutdown, important state changes). Think of it as a “verbose” mode. Example: logger.info("Connection pool initialised with {} threads", poolSize);
3. WARN for potential issues that do not stop the application from functioning. Use it when you fall back to a default, or when a deprecated API is called. The system carries on without a problem. logger.warn("Configuration key 'max.retries' missing – defaulting to 3");
4. ERROR for application errors that still allow the application to hobble along. These indicate a real problem – an exception caught, a service call failed, but the operation can be retried or the user can continue. Like a missing admin‑supplied config value that falls back to a hard‑coded default. logger.error("Unable to reach inventory service, using cached prices", ex);
5. FATAL for critical failures, after which the application quits abnormally. After logging a fatal message, the system is considered dead. For example, in Log4j2: logger.fatal("Database connection pool exhausted, aborting.");

Additionally, two special levels exist with obvious meanings:

1. ALL – enables every log statement, regardless of its level. Handy for temporarily debugging a specific component, but never for production.
2. OFF – disables all logging output. Use it to silence a noisy logger completely.