I have Standalone Flink Cluster. When I stop the process on Task Manager, classes that were loaded as ChildFirst are not removed. After several start/stop repetitions, the metaspace exceeds the maximum value and throws OutOfMemory Metaspace.
I run servals application in one task manager. Application are different so a cannot add the JAR file to /lib folder. Restarting TaskManager after severals operation is bothersome.

I have the following class:

i have a project in Flink 1.11.3 where i can observe some Backpressure on one of the Pipelines:

Team ,

I am working with an Apache Flink application that reads data streams from Kafka. The application processes the streams, creating objects from the data and calculating the processing time by marking the start time at object creation and the end time after performing a series of transformations including keyBy and some calculations.