Antithesis Stress and Deterministic Workloads (for L1 chain)

[parthshah1]

Hello team, for making most out of antithesis, where features like time travel in simulation and determinism can lead us to test for really complex and peculiar behavior. Rather than blackbox fuzzing (which is important too), but focusing on consensus/state related bugs will lead to more critical results.

Some of them include (not exhaustive):

1. Extreme gas messages: Messages simulating high gas messages, tracking basefee spikes.
2. F3/EC Divergence: Adding forky chains with heavy weight than canonical and asserting that it doesn't corrupt stateroot.
3. Double Spending: Miner equivocation. Remove slash filter from nodes to stress this behavior.
4. Consensus Split: Divergence between the ParentBaseFee in the block header and the validator’s calculated fee due to rounding/overflow.
5. Mempool message selection logic: Conflicting messages, message chains and assert which once should get picked across nodes.
6. FVM stress: Adding smart contracts/expensive VM messages that exhaust stacks and consume lots of gas.
7. Transaction with invalid/forged/mangled signatures, batch transactions.
8. Large logs emission: To test for log spam, OOM issues.
9. Miner withhelding the block: An adversary mines blocks privately and selectively releases them to invalidate the work of honest miners.

Looking forward from more inputs from the teams on what should we focus more on.

Thanks.