memory leaks from libgrpc and libgrpc++ when using opentelemetry-cpp

[farzinlize]

hey guys, I previously asked the community about the memory leak reports that I got while using http exporters. I submitted an issue #3642 which was answered. now when I'm using grpc exporters, I faced the same problem. here is my valgrind report:

==2076== HEAP SUMMARY:
==2076==     in use at exit: 4,600 bytes in 6 blocks
==2076==   total heap usage: 5,047 allocs, 5,041 frees, 784,390 bytes allocated
==2076== 
==2076== 8 bytes in 1 blocks are still reachable in loss record 1 of 6
==2076==    at 0x4846FA3: operator new(unsigned long) (in /usr/libexec/valgrind/vgpreload_memcheck-amd64-linux.so)
==2076==    by 0x573E6F8: ??? (in /usr/lib/x86_64-linux-gnu/libgrpc.so.29.0.0)
==2076==    by 0x400571E: call_init.part.0 (dl-init.c:74)
==2076==    by 0x4005823: call_init (dl-init.c:120)
==2076==    by 0x4005823: _dl_init (dl-init.c:121)
==2076==    by 0x401F59F: ??? (in /usr/lib/x86_64-linux-gnu/ld-linux-x86-64.so.2)
==2076== 
==2076== 8 bytes in 1 blocks are still reachable in loss record 2 of 6
==2076==    at 0x4846FA3: operator new(unsigned long) (in /usr/libexec/valgrind/vgpreload_memcheck-amd64-linux.so)
==2076==    by 0x4BF7BD7: ??? (in /usr/lib/x86_64-linux-gnu/libgrpc++.so.1.51.1)
==2076==    by 0x4BF7D33: ??? (in /usr/lib/x86_64-linux-gnu/libgrpc++.so.1.51.1)
==2076==    by 0x400571E: call_init.part.0 (dl-init.c:74)
==2076==    by 0x4005823: call_init (dl-init.c:120)
==2076==    by 0x4005823: _dl_init (dl-init.c:121)
==2076==    by 0x401F59F: ??? (in /usr/lib/x86_64-linux-gnu/ld-linux-x86-64.so.2)
==2076== 
==2076== 8 bytes in 1 blocks are still reachable in loss record 3 of 6
==2076==    at 0x4846FA3: operator new(unsigned long) (in /usr/libexec/valgrind/vgpreload_memcheck-amd64-linux.so)
==2076==    by 0x4BF7C0C: ??? (in /usr/lib/x86_64-linux-gnu/libgrpc++.so.1.51.1)
==2076==    by 0x4BF7D33: ??? (in /usr/lib/x86_64-linux-gnu/libgrpc++.so.1.51.1)
==2076==    by 0x400571E: call_init.part.0 (dl-init.c:74)
==2076==    by 0x4005823: call_init (dl-init.c:120)
==2076==    by 0x4005823: _dl_init (dl-init.c:121)
==2076==    by 0x401F59F: ??? (in /usr/lib/x86_64-linux-gnu/ld-linux-x86-64.so.2)
==2076== 
==2076== 8 bytes in 1 blocks are still reachable in loss record 4 of 6
==2076==    at 0x4846FA3: operator new(unsigned long) (in /usr/libexec/valgrind/vgpreload_memcheck-amd64-linux.so)
==2076==    by 0x4BF7E5C: ??? (in /usr/lib/x86_64-linux-gnu/libgrpc++.so.1.51.1)
==2076==    by 0x400571E: call_init.part.0 (dl-init.c:74)
==2076==    by 0x4005823: call_init (dl-init.c:120)
==2076==    by 0x4005823: _dl_init (dl-init.c:121)
==2076==    by 0x401F59F: ??? (in /usr/lib/x86_64-linux-gnu/ld-linux-x86-64.so.2)
==2076== 
==2076== 56 bytes in 1 blocks are still reachable in loss record 5 of 6
==2076==    at 0x4846FA3: operator new(unsigned long) (in /usr/libexec/valgrind/vgpreload_memcheck-amd64-linux.so)
==2076==    by 0x573F52D: ??? (in /usr/lib/x86_64-linux-gnu/libgrpc.so.29.0.0)
==2076==    by 0x400571E: call_init.part.0 (dl-init.c:74)
==2076==    by 0x4005823: call_init (dl-init.c:120)
==2076==    by 0x4005823: _dl_init (dl-init.c:121)
==2076==    by 0x401F59F: ??? (in /usr/lib/x86_64-linux-gnu/ld-linux-x86-64.so.2)
==2076== 
==2076== 4,512 bytes in 1 blocks are still reachable in loss record 6 of 6
==2076==    at 0x48485C3: operator new[](unsigned long) (in /usr/libexec/valgrind/vgpreload_memcheck-amd64-linux.so)
==2076==    by 0x573D599: ??? (in /usr/lib/x86_64-linux-gnu/libgrpc.so.29.0.0)
==2076==    by 0x573FFC8: ??? (in /usr/lib/x86_64-linux-gnu/libgrpc.so.29.0.0)
==2076==    by 0x400571E: call_init.part.0 (dl-init.c:74)
==2076==    by 0x4005823: call_init (dl-init.c:120)
==2076==    by 0x4005823: _dl_init (dl-init.c:121)
==2076==    by 0x401F59F: ??? (in /usr/lib/x86_64-linux-gnu/ld-linux-x86-64.so.2)

previously I resolved my issue by using these lines on shutdown:

/* shutdown protobuf */
google::protobuf::ShutdownProtobufLibrary();
/* shutdown gnutls */
gnutls_global_deinit();

now I'm asking what function I should call in libgrpc and libgrpc++ to avoid those memory leak reports above?