Structured logging represents each log event as machine-parseable fields instead of unstructured prose. Programs can then extract values such as level, message, and time for dashboards, error counts, and request-duration measurements. section 4.5
A practical structured logging implementation should:
- Encode a record such as a level-and-message structure with encoding/json and Marshal. section 4.5
- Export structure fields so the JSON package can read and write them. section 4.5
- Prefer a structured logger for concurrent production services rather than the basic log package. section 8.6.3
A structured warning
A JSON event can record a warning that a platform was not scaled up for a request, while preserving fields that programs can parse. section 4.5
{
"time": "2022-31-10 23:06:30.148845Z",
"level": "warning",
"message": "platform not scaled up for request"
}
Inspect a concrete JSON log record with time, level, and message fields. section 4.5
