Railway’s log viewer currently treats application logs as plain text streams. While this works for simple logging, it limits the usefulness of structured logging frameworks commonly used in modern applications such as Winston, Pino, and Bunyan.

Many production systems emit logs as structured JSON objects containing metadata fields such as request IDs, account IDs, error objects, or service context. For example in TS:

logger.warn('Failed to log a whole action.', {

someData,

anotherValue,

foo,

bar,

error,

});

When this get to logs I can only see 'Failed to log a whole action.' I miss context and stack-traces as a result. It would be nice if it custom fields got added to the raw or attributes section in logging.

This could be user error - but other logging solutions I used before porting to Railway (Splunk) seemed to pick this up natively without additional config.

Thanks all!