I have a deployed Apache Ignite in Kubernetes cluster with SpringBoot microservice as the client node for Apache Ignite.
In one of my testing environment, I have some failing client nodes that are constantly restarting themselves and rejoin the Apache Ignite topology every time those nodes restarted.
However, I have found that when the above situation happens, Apache Ignite is going to show OutOfMemory error and the haywire will goes haywire.
The HeapDump analysis shows that the class that has highest heap memory usage is the class that perform the TCPDiscovery and the code shows that Apache Ignite is adding all connections to a list for each nodes joining.
I would like to know: