[CORE-643] fix: handle exec and fork from non-leader threads

[RonFed]

This PR contains 2 main improvements:

• avoid reporting to user space on failed execs.
• properly handle exec events performed by not the main thread. This was an edge case we didn't properly handled - since the pid as seen in the exec entry is not the actual one used in this case. Since after an exec by a non-leader thread - all the other threads are terminated, and the new execution will be under the thread group leader - tgid.
• properly handle fork events done by not the main thread. The fork probe is updated to look the parent tgid and not the parent pid.

additional improvemts

• add an ongoing_execs map to only correlate the exec exit and exec entry probes. - The existing tracked_pids maps are only updated at the exec return probe - to avoid adding failed exec events to them.
• add test cases for a failed exec and for an exec done by a non leader thread.
• add assertions for empty maps after these use cases - to make sure no leakage in the eBPF maps.