MongoDB can be used for log centralization due to its asynchronous insert capabilities, which prevent slowing down the user experience. The use of capped collections ensures old log data is automatically LRU'd out, preventing disk space issues and the need for log archival scripts. MongoDB's fast performance, combined with document-oriented JSON format, makes it a suitable choice for logging applications. Additionally, the profiler works similarly to a log collection, storing profile timings in a collection that can be queried.