Timing attacks do not run in parallel
Problem
Timing attacks that are run in parallel could potentially cause false positives for each other. For example, if two timing attacks that cause a lock on a database are run at the same time, then if one is successful then both attacks may raise a finding, one of which would be a false positive.
Proposal
Timing attacks should never run in parallel.
A higher level discussion with the team should take place about what architecture is best to run attacks in parallel. For example, it may be best to run attacks in parallel, or there could be a workers queue, etc.