These are possible key performance indicators (KPIs) to benchmark the optimization of our workflow:
- number and size of memory allocations
- overhead in horizontal scaling (time of request for one single point with 1 or multiple threads/processes/nodes)
- speed-up factor for a given test dataset (time for 1,2,n threads/processes/nodes)
- time to first response for a single point (container spin-up time)