Pinpoint pinpoint Provide Logging context analysis tools to find important information in vast amounts of logging Pinpoint is designed to allow log correlation using various mechanisms built into log4j (Mapped Diagnostic Contexts, Nested Diagostic Contexts, event properties). For example, using MDC a thread can set a 'userId' key so that each LoggingEvent generated by that Thread is tagged with the same set of MDC values. This allows LoggingEvents to be correlated together later. When millions of logging events are involved, finding specific sets of events is difficult, and often involves command-line tools (grep) to filter. Unfortunately these tools are not LoggingEvent-aware and can strip out important information. Pinpoint uses Apache Lucene to index the Logging content and provide an easily searchable context, backed up with a simple binary event data-store so that the full, rich, event detail of the LoggingEvent is retained (this reduces the need to store the entire binary content in a Lucene index, which is not very efficient during high volume indexing). The project is designed to provide both command-line and web-based tools allowing easy navigation to find relevant events in the vast ocean of production events. If this sounds a lot like the commercial Splunk application, that's because it's inspired in part by it and by the authors need to drill through events to find user-specific application usage behaviour. 2007-8-5 idle Paul Smith f33fdcb4bbe365a3295535d7112499923feaa4e3 Java