Description

A multiplexer processing cycle exceeded the configured notification threshold. Common causes include concurrent garbage collections (enable JVM garbage collection logging to investigate); overly general topic selectors that must be tested against many topics (prefer topic selectors with more specific prefix paths); subscription processing for many sessions (if more CPU cores are available, consider increasing the number of multiplexers); more multiplexers configured than available CPU cores (reduce the number of multiplexers). If the reported operation count is low, reducing the multiplexer monitoring period threshold will provide further diagnostics.