Investigate cause of variation in fuel consumption in random.js integration test

[jeffcharles]

In this run of the CI workflow, I observed the following error message:

thread 'test_same_module_outputs_different_random_result' panicked at 'fuel_consumed (130653) was not within 10.00% of the target_fuel value (100543)', crates/cli/tests/integration_test.rs:163:5

It's not unexpected that the fuel consumed would variable between test runs for the random script since the inputs will be different from test run to test run. However, it is unexpected that it could be 30% different between runs. We should investigate to better explain the source of the variation.