Currently we create a fixed number of evms (based on a config param) for eth_call (past and pending) and debug_traceTransaction. We should make it so the config only sets a baseline of threads to be created on startup, but that can scale up to some limit based on the volume of calls (and scales back as the volume decreases).
|
max(config.executor_call_present_evms.unwrap_or(config.executor_evms / 2), 1), |