Purpose of logging feature
Especially in higher volume environments it can be of interest to have detailed information about the pdfToolbox processes as they happen. Suitable logging data can be a good foundation both for monitoring processing of PDF files as well as for post mortem analyses or analyses aiming to improve overall behavior of the whole processing system.
The logging feature that forms a part of pdfToolbox from version 10 onwards provides logging information in the form of JSON files that get written immediately at the following steps during an invocation of pdfToolbox:
- When pdfToolbox is launched; only very basic data about the PDF file to be produced – such as file name – is available at this time, plus some information about the environment or the command line parameters
- When pdfToolbox initiates processing of a PDF file; additional data about the PDF to be processed is provided – only aspects though that do not require no 'deeper' analysis in order not to hamper runtime behavior – such as file size or PDF version or size of the first page; in addition all variables – as evaluated at the beginning of processing – are included as well.
- When pdfToolbox completes processing of a PDF file; at this stage various aspects of overview information is readily available, such as colors, fonts or ICC profiles used, output intents, conformance with PDF/X or other PDF standards.
Command line and Server versus Desktop
Logging is available in command line, Server and Desktop versions. It is not expected that logging on desktop is normally being used. Instead, on Desktop activating logging will help understand how logging works, and often makes it easier to design and test use of logging data outside of an automated production setup.
How to process logging data
Research has shown that most callas software customers who operate higher volume environments prefer any logging data to be provided rather as many small files using JSON over using other formats or a small number of large files. Beyond the way the logging data is provided, no assumptions are being made how these data are processed.