Reduce default memory allocation to the java process

[amahussein]

Fixes #1406

This pull request introduces several changes to improve the handling of JVM heap size and thread calculations in the spark_rapids_tools module. The most important changes include updating the method for calculating JVM heap size.

This change aims at avoiding allocating memory by default that would trigger the OOM-killer

• use available memory instead of total.
• cap the xmx to 32 GB
• cap the max number of threads to 8

Enhancements to JVM heap size and thread calculations:

• user_tools/src/spark_rapids_tools/cmdli/argprocessor.py: Updated the method to calculate JVM heap size using Utilities.calculate_jvm_max_heap_in_gb() instead of Utilities.get_system_memory_in_gb(). Increased the minimum heap size per thread from 6 GB to 8 GB.

• user_tools/src/spark_rapids_tools/utils/util.py: Added class variables min_jvm_xmx, max_jvm_xmx, and max_tools_threads to set limits on JVM heap size and the number of threads.

Method renaming for clarity:

• user_tools/src/spark_rapids_tools/utils/util.py: Renamed get_system_memory_in_gb() to calculate_jvm_max_heap_in_gb() and updated the method to calculate the maximum heap size based on available system memory, capping it between 8 GB and 32 GB.

• user_tools/src/spark_rapids_tools/utils/util.py: Renamed get_max_jvm_threads() to calculate_max_tools_threads() and updated the method to calculate the maximum number of threads based on physical cores, capping it at 8 threads.

• user_tools/src/spark_rapids_tools/utils/util.py: Updated the adjust_tools_resources method to use the new calculate_max_tools_threads() method for determining the maximum number of threads.

[parthosa]

Thanks @amahussein.