Consider better `timeout` handling to optimize all parts of Hop
shanefontaine opened this issue · 0 comments
shanefontaine commented
Consider a more robust mental model for how long a timeout should be for various situations (RPC calls, web2 service calls, etc.). Could possibly dramatically increase TPS.
Possible Implementation
Strict timeout for tx lifecycle stuff that affects TPS and less-strict for non-tx lifecycle
Considerations
- Different RPC providers have different speeds and handling them all might be impossible.
- A long timeout (seconds) could block the bonder and we should not allow that.
- Linea’s (Infura) RPC took 7s and 9s for two respective txs during their quest…