throttle "refund" for gas/sec should be based on _consumed_ not _used_

[david-bakin-sl]

Background

We have two different measures of gas eaten by a contract:

• consumed - actual amount computed by the EVM (+ our charges for system contract methods + intrinsic minimum)
• used - amount charged caller which is consumed + the 80% tariff for overestimating the gas limit on submitting a transaction

The gas throttle limiting how many contract executions we run per second is done based on the gas limit - because the transaction hasn't run yet. But after the transaction is run and we know how much gas was actually used, we "refund" the excess back to the throttle so that we can accept more work in that period.

This happens in DispatchUsageManager.leakUnusedGas.

And if I'm reading that correctly, it's based on gas used. But since this is a performance limit (on the network) it should be based instead on gas consumed.

Acceptance Criteria

1. Difference in amount refunded to throttle based on transactions with gas limit, gas consumed, and gas used (based on overestimates of gas limit).
2. Performance testing showing that after this change you get more contract execution tx/sec accepted and executed even though those contract txs together would hit the 15M throttle based on gas limit, or based on total gas used, but not with total gas consumed.

Dependencies

No response

Definition of Ready (DoR) Checklist

• Clear acceptance criteria
• Clear and detailed description
• Dependencies identified
• Links to documentation
• Should be completable in 2-3 Days
• Initial draft of Low-level design document
• At least high level test plan
• Groomed/Estimated

Definition of Done (DoD) Checklist

• Acceptance Criteria complete
• No Codacy issues greater than minor (in new code)
• JavaDocs updated/created
• Code commented
• Unit tests created/updated
• 80% test code coverage (in new code)
• Happy Path and major negative cases in HAPI tests as applicable