Logahawk provides more Severity levels than most Logging frameworks. So what is the best way to use these different levels?

• Debug - Used by developers to create log statements for developers. Debug statements are usually used in two ways: used often to allow a clear view into what the system is doing at any point, or used only when a developer needs to communicate something to another developer.
• Info - The most common Severity level. This should be used to record any interesting event that occurs.
• Alert - Used to record interesting events that are more significant than Info statements but do not indicate any sort of problem. This can be used to log make statements standout from the sea of Info statements. Long running applications are more likely to make use of this Severity.
• Warn - Used to warn of unexpected but recoverable or inconsisquential behavior. The decision of whether something is a warning or error depends on how it affects the operation in question.
• Error - Used record errors, such as Exceptions. Depending on the application this can be used for errors that halt the entire application or errors in a particular operation that the application is managing. See the description of Fatal and Panic.
• Fatal - Indicates that an operation failed entirely or the application itself failed. This should be used for errors more significant than Error. See the description of Panic.
• Panic - Indicates an unrecoverable error, one that prevents the application from functioning. This should be used when the application must shutdown. This Severity is based in the idea of the panic() function in UNIX.

A good opportunity to log is at the beginning and end of significant operations. This includes program start and finish.

For server applications that control many sub-operations, either use Alerts to indicate the start and stop of operations, use custom LogMeta's to tag log statements for each operation, or use separate log files. (Custom LoggerRegistry can help create dynamic Loggers for each operation.)