Unbounded disk space usage by capture feature

Description

The capture feature of Tracee works by writing artifacts to the disk whenever some operation happens.
Some of the capture options might be exposed to huge loads in some environment, results the creation of many artifacts files on the disk.
However, disk is a resource of the operating system, and using it without bounds might result faults and unreasonable expenses for the user.
We should be able to address the disk usage issue in some manner.

Solution Options in My POW

Option 1: Stream Artifacts

If we stream the artifacts, than from that point on the responsibility for resources usage is on the other side of the stream.
In this case there is nothing we can do about it if we look at it naively.

Option 2: Bound Memory and Prioritize Artifacts

If we do use the memory with Tracee, we should allow the user to configure the maximum amount of disk Tracee can use.
If we implement it naively, it means just drop new captures after some load is captured.
In more practical thinking, I think we should prioritize new captures, and expose a way to prioritize captures in general.
We have the rules that determine if something is interesting or not, so I think that we can use rules detections to understand if some capture is important and avoid deleting it.

The issue does not name specific files or tests. Start by tracing the capture feature’s artifact-writing path and reviewing how capture options and rules detections determine artifacts. Define the disk limit and prioritization behavior, then verify that sustained capture load respects the configured bound without losing higher-priority artifacts.