Both system log and various state status entries expect information in a well defined format that can be properly handled in backend system processing.
General indication of the severity of the message, using standard categorizations:
| Level | Label | Description |
|---|---|---|
| 100 | DEBUG | Stream of consciousness useful for detailed debugging. |
| 200 | INFO | Just information about all that is well in the world. |
| 300 | NOTICE | Normal but significant events, such as start up, shut down, or a configuration change. |
| 400 | WARNING | Something is not right that should be investigated. |
| 500 | ERROR | Something bad that needs immediate attention. |
The entry category is a dot-separated string providing a semantic hierarchy for a given message. Canonical categories correspond to expected values for log and status entries. These are designed to be automatically categorized and processed by backend systems.
The entry message is a single sentence of human-readable output describing what went wrong. This will likely be directly exposed to a system operator to give them high-level information on how to triage, diagnose, or rectify the situation. Exact messages are not prescribed, although some specific categories will have guiding principles (e.g. prepend point message with “Point pointname”).
The entry detail is use for specifically diagnosing the error. E.g. a complete stack-trace or parsing message that details exactly where the error occurred. This would be used for detailed debugging by a domain expert.