Files
kserve/test/benchmark/sk_benchmark.yaml
Dan SunandGitHub 941a38eb1b Remove default queue proxy resource limit and Add KFServing benchmarking (#894)
* Add benchmark script

* Add sklearn bechmarking

* Add tf benchmark

* Add perf result for autoscaler and activator

* Add HPA experiments

* Address comments

* Bump up tf flower example cpu limit

* Bump cpu limit for tf benchmark

* Update hpa result for tf

* Fix hpa benchmark
2020-07-06 06:44:52 -07:00

38 lines
862 B
YAML

apiVersion: batch/v1
kind: Job
metadata:
generateName: sklearn-load-test
spec:
backoffLimit: 6
parallelism: 1
template:
metadata:
annotations:
sidecar.istio.io/inject: "false"
spec:
containers:
- args:
- vegeta -cpus=1 attack -duration=10m -rate=5/1s -targets=/var/vegeta/cfg
| vegeta report -type=text
command:
- sh
- -c
image: peterevans/vegeta:latest
imagePullPolicy: Always
name: vegeta
resources:
requests:
cpu: "1"
memory: 1Gi
volumeMounts:
- mountPath: /var/vegeta
name: vegeta-cfg
restartPolicy: Never
volumes:
- configMap:
defaultMode: 420
name: vegeta-cfg
name: vegeta-cfg
---