// TestServiceAllocPerformance measure the latency to create N services with a parallelism of K // using the old and the new ClusterIP allocators. // The test is skipped to run on CI and is left to execute manually to check for possible regressions. // The current results with 100 works and 15k services on a (n2-standard-48) vCPU: 48 RAM: 192 GB are: // legacy perf_test.go:139: [RESULT] Duration 1m9.646167533s: [quantile:0.5 value:0.462886801 quantile:0.9 value:0.496662838 quantile:0.99 value:0.725845905] // new perf_test.go:139: [RESULT] Duration 2m12.900694343s: [quantile:0.5 value:0.481814448 quantile:0.9 value:1.3867615469999999 quantile:0.99 value:1.888190671] func TestServiceAllocPerformance(t *testing.T) { … }