[Data] Add ClusterUtil dataclass for ResourceUtilizationGauge

[bveeramani]

ExecutionResources can't represent object store memory and memory utilization correctly because it rounds values to the nearest integer. So, the utilization for those resources is always either 0 or 1.

To ensure we can express the full [0, 1] range of resource utilization, this PR introduces a new ClusterUtil dataclass rather than re-using ExecutionResources.

[gemini-code-assist]

Code Review

This pull request introduces a new ClusterUtil dataclass to correctly handle floating-point resource utilization values, which is a good improvement. The implementation is sound. I've added a couple of suggestions to refactor the validation logic in the new dataclass to improve its maintainability by reducing code duplication.

[gemini-code-assist]

To support iterating over dataclass fields for validation in __post_init__, please also import fields here. This will be used in the refactoring of __post_init__.

Suggested change

	from dataclasses import dataclass
	from dataclasses import dataclass, fields

[gemini-code-assist]

To improve maintainability and reduce code duplication, you can iterate over the dataclass fields to perform the validation. This approach will automatically include any new fields added to the dataclass in the future, making the code more robust.

Suggested change

	def __post_init__(self):
	# If we overcommit tasks, the logical utilization can exceed 1.0.
	assert math.isfinite(self.cpu) and 0 <= self.cpu
	assert math.isfinite(self.gpu) and 0 <= self.gpu
	assert math.isfinite(self.memory) and 0 <= self.memory
	assert math.isfinite(self.object_store_memory) and 0 <= self.object_store_memory
	def __post_init__(self):
	# If we overcommit tasks, the logical utilization can exceed 1.0.
	for f in fields(self):
	value = getattr(self, f.name)
	assert math.isfinite(value) and value >= 0, f"Invalid value for {f.name}: {value}"

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, have a team admin enable autofix in the Cursor dashboard.}