logging

Creek Observability Logging

Wrapper around Slf4J aimed at encouraging more structured logging.

More structured logging is more easily ingested by log aggregation and visualization tools, such as Splunk, the ELK stack, etc.

For example, the code snippet below logs events with structured metrics:

class SomeType {

    private static final StructuredLogger LOGGER = StructuredLoggerFactory.logger(SomeType.class);
    
    void someMethod(int attempt) {
        final Instant start = Instant.now();
        
        LOGGER.trace("Starting", log -> log.ns("my.app").with("attempt", attempt));
        doStuff();
        LOGGER.info("Did stuff", log -> log.ns("my.app").with("attempt", attempt).with("took", Duration.between(start, Instant.now())));
    }
}

Depending on the logging framework used and how it is configured, this might result in a log line such as:

{
  "timestamp" : "2022-03-03T14:20:37.414+0000",
  "level" : "INFO",
  "loggerName" : "org.acme.SomeType",
  "message" : "{\"message\":\"Did stuff\",\"my.app\":{\"took\":\"PT0.003938S\",\"attempt\":2}}",
  "threadId" : 1,
  "thread" : "main",
  "env" : "PROD",
  "aggregate" : "PRICING",
  "service" : "BLOOMBERG",
  "version" : "2.1.9",
  "host" : "aws.server1"
}

The message field contains a nested JSON document containing the message and metrics being logged.