Status
Ready to be implemented. Potential good use case for first contribution.
Description
Maiko doesn't have any benches, so we don't have any good visibility over basic performance. Let's start with a minimum, Criterion-based benches.
Scenario 1 - actor to actor message transport. Just two actors, doing nothing nothing else than sending/receiving events. That tests measures maximum real throughput. so the actual cost of Actor->Broker->Actor jump. Time as a unit.
Scenario 2 - measure the impact of number of actors for single message delivery time. Assume that 2 actors delivery is 1, how proportionally slower it is for 3, 4, 10, etc.
You are welcome to propose other benches (i.e. topic impact) or modify scenario 2 if needed.
Tasks / Acceptance criteria
Status
Ready to be implemented. Potential good use case for first contribution.
Description
Maiko doesn't have any benches, so we don't have any good visibility over basic performance. Let's start with a minimum, Criterion-based benches.
Scenario 1 - actor to actor message transport. Just two actors, doing nothing nothing else than sending/receiving events. That tests measures maximum real throughput. so the actual cost of Actor->Broker->Actor jump. Time as a unit.
Scenario 2 - measure the impact of number of actors for single message delivery time. Assume that 2 actors delivery is 1, how proportionally slower it is for 3, 4, 10, etc.
You are welcome to propose other benches (i.e. topic impact) or modify scenario 2 if needed.
Tasks / Acceptance criteria