Each detector instance maintains an internal 10 KB memory buffer that it uses to maintain debug messages. These are written to the dedicated log file when a debug event (an unexpected resource status report) occurs. The buffer is circular, so when a new message would cause the buffer to overflow, it will instead overwrite one or more of the oldest stored messages. This continues until a debug event causes the entire buffer to be written to the log file, at which time the buffer is reset. Therefore, only the latest messages will be logged. Some older data may be lost, depending on when a debug event occurs.
When the buffer contents are flushed to the log file, the messages are written in chronological order. It is important to note that the latest message in the buffer was generated by the debug event, but earlier messages in the buffer may or may not be related. Therefore, when you examine the detector log, you may find a series of benign messages, followed by a debug event message, followed by more benign messages, followed by another debug event message, etc. This means you must check the timestamp and text of every detector debug message to see if it pertains to the problem you're trying to resolve.
Every message that the detector can generate has an assigned detail level, where higher detail levels produce more debugging information. At runtime, only those messages with a detail level less than or equal to the detector's current log level will be added to the internal buffer. The default value of each detector log level is set to 1.
Note
The amount of information produced by a specific log level depends on the detector resource type. A high log level for one detector may produce less output than a low log level for another detector.
