I have know this words in the Stream logs from Google Cloud to Splunk reference architecture : A single vCPU in a Splunk Dataflow pipeline can generally process 3k EPS
How is 3k EPS calculated? where is all machine types' performance benchmark?
The 3k EPS per vCPU is a general benchmark, based on real-world tests using typical workloads and machine types (like n1-standard-1).
For higher throughput, scale horizontally (more workers) or vertically (larger machine types).
Google does not publish detailed EPS benchmarks for each machine type, but you can benchmark them yourself using Pub/Sub and Dataflow metrics.