github/kubernetes
1
0
Fork 0
mirror of https://github.com/k3s-io/kubernetes.git synced 2025-08-01 15:58:37 +00:00
Code Issues Projects Releases Wiki Activity

scheduler_perf: add thresholds to DRA test cases

Browse Source 
They were enabled yesterday and executed seven times, with results that (so
far) seem to be fairly stable with just one run that was slower across the
board.

The links in the YAML can be used to navigate to each test case quickly. The
thresholds were chose with a 20% security margin below what seems to be a
common result.
This commit is contained in:
Patrick Ohly 2025-01-31 19:41:17 +01:00
parent b4f902f037
commit e2ff03486d
1 changed files with 16 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
test/integration/scheduler_perf/dra/performance-config.yaml
View File

@ -93,6 +93,8 @@
featureGates:
SchedulerQueueingHints: false
labels: [performance]
# https://perf-dash.k8s.io/#/?jobname=scheduler-perf-benchmark&metriccategoryname=Scheduler&metricname=BenchmarkPerfScheduling&Metric=SchedulingThroughput&Name=BenchmarkPerfScheduling%2FSchedulingWithResourceClaimTemplate%2F5000pods_500nodes%2Ftest&event=not%20applicable&extension_point=not%20applicable&plugin=not%20applicable&result=not%20applicable
threshold: 40 # typically above 51
params:
nodesWithDRA: 500
nodesWithoutDRA: 0
@ -103,6 +105,8 @@
featureGates:
SchedulerQueueingHints: true
labels: [performance]
# https://perf-dash.k8s.io/#/?jobname=scheduler-perf-benchmark&metriccategoryname=Scheduler&metricname=BenchmarkPerfScheduling&Metric=SchedulingThroughput&Name=BenchmarkPerfScheduling%2FSchedulingWithResourceClaimTemplate%2F5000pods_500nodes_QueueingHintsEnabled%2Ftest&event=not%20applicable&extension_point=not%20applicable&plugin=not%20applicable&result=not%20applicable
threshold: 56 # typically above 70
params:
nodesWithDRA: 500
nodesWithoutDRA: 0
@ -191,6 +195,8 @@
featureGates:
SchedulerQueueingHints: false
labels: [performance]
# https://perf-dash.k8s.io/#/?jobname=scheduler-perf-benchmark&metriccategoryname=Scheduler&metricname=BenchmarkPerfScheduling&Metric=SchedulingThroughput&Name=BenchmarkPerfScheduling%2FSteadyStateClusterResourceClaimTemplate%2Fempty_500nodes%2Ftest&event=not%20applicable&extension_point=not%20applicable&plugin=not%20applicable&result=not%20applicable
threshold: 64 # typically above 81
params:
nodesWithDRA: 500
nodesWithoutDRA: 0
@ -201,6 +207,8 @@
featureGates:
SchedulerQueueingHints: true
labels: [performance]
# https://perf-dash.k8s.io/#/?jobname=scheduler-perf-benchmark&metriccategoryname=Scheduler&metricname=BenchmarkPerfScheduling&Metric=SchedulingThroughput&Name=BenchmarkPerfScheduling%2FSteadyStateClusterResourceClaimTemplate%2Fempty_500nodes_QueueingHintsEnabled%2Ftest&event=not%20applicable&extension_point=not%20applicable&plugin=not%20applicable&result=not%20applicable
threshold: 64 # typically above 81
params:
nodesWithDRA: 500
nodesWithoutDRA: 0
@ -226,6 +234,8 @@
featureGates:
SchedulerQueueingHints: false
labels: [performance]
# https://perf-dash.k8s.io/#/?jobname=scheduler-perf-benchmark&metriccategoryname=Scheduler&metricname=BenchmarkPerfScheduling&Metric=SchedulingThroughput&Name=BenchmarkPerfScheduling%2FSteadyStateClusterResourceClaimTemplate%2Fhalf_500nodes%2Ftest&event=not%20applicable&extension_point=not%20applicable&plugin=not%20applicable&result=not%20applicable
threshold: 70 # typically over 78
params:
nodesWithDRA: 500
nodesWithoutDRA: 0
@ -236,6 +246,8 @@
featureGates:
SchedulerQueueingHints: true
labels: [performance]
# https://perf-dash.k8s.io/#/?jobname=scheduler-perf-benchmark&metriccategoryname=Scheduler&metricname=BenchmarkPerfScheduling&Metric=SchedulingThroughput&Name=BenchmarkPerfScheduling%2FSteadyStateClusterResourceClaimTemplate%2Fhalf_500nodes_QueueingHintsEnabled%2Ftest&event=not%20applicable&extension_point=not%20applicable&plugin=not%20applicable&result=not%20applicable
threshold: 60 # typically over 75
params:
nodesWithDRA: 500
nodesWithoutDRA: 0
@ -261,6 +273,8 @@
featureGates:
SchedulerQueueingHints: false
labels: [performance]
# https://perf-dash.k8s.io/#/?jobname=scheduler-perf-benchmark&metriccategoryname=Scheduler&metricname=BenchmarkPerfScheduling&Metric=SchedulingThroughput&Name=BenchmarkPerfScheduling%2FSteadyStateClusterResourceClaimTemplate%2Ffull_500nodes%2Ftest&event=not%20applicable&extension_point=not%20applicable&plugin=not%20applicable&result=not%20applicable
threshold: 60 # typically over 75
params:
nodesWithDRA: 500
nodesWithoutDRA: 0
@ -271,6 +285,8 @@
featureGates:
SchedulerQueueingHints: true
labels: [performance]
# https://perf-dash.k8s.io/#/?jobname=scheduler-perf-benchmark&metriccategoryname=Scheduler&metricname=BenchmarkPerfScheduling&Metric=SchedulingThroughput&Name=BenchmarkPerfScheduling%2FSteadyStateClusterResourceClaimTemplate%2Ffull_500nodes_QueueingHintsEnabled%2Ftest&event=not%20applicable&extension_point=not%20applicable&plugin=not%20applicable&result=not%20applicable
threshold: 60 # typically over 75
params:
nodesWithDRA: 500
nodesWithoutDRA: 0