h2load

Benchdog: h2load

A containerized HTTP/2 benchmark based on h2load.

Running the server in Kubernetes

Server namespace:

kubectl create namespace benchdog-server
kubectl config set-context --current --namespace benchdog-server
kubectl apply -f server/

Running the client in Kubernetes

Client namespace:

kubectl create namespace benchdog-client
kubectl config set-context --current --namespace benchdog-client
kubectl run -it --rm --env BENCHDOG_HOST=<host> --image quay.io/skupper/benchdog-h2load-client h2load-client

Testing with Skupper

Server namespace:

kubectl config set-context --current --namespace benchdog-server
skupper init
skupper token create ~/token.yaml
skupper expose deployment/h2load-server --port 58080

Client namespace:

kubectl config set-context --current --namespace benchdog-client
skupper init
skupper link create ~/token.yaml

Once the h2load-server service appears in the client namespace, you can run the client with --env BENCHDOG_HOST=h2load-server.

Client namespace:

kubectl run -it --rm --env BENCHDOG_HOST=h2load-server --image quay.io/skupper/benchdog-h2load-client h2load-client

Example commands

kubectl run -it --rm --env BENCHDOG_HOST=h2load-server --env BENCHDOG_SCENARIOS=100:100,100:200,100:300,100:400,100:500 --env BENCHDOG_ITERATIONS=3 --env BENCHDOG_DURATION=60 --image quay.io/skupper/benchdog-h2load-client h2load-client

kubectl run -it --rm --env BENCHDOG_HOST=h2load-server --env BENCHDOG_SCENARIOS=100:100,500:100 --env BENCHDOG_ITERATIONS=3 --env BENCHDOG_DURATION=60 --image quay.io/skupper/benchdog-h2load-client h2load-client

kubectl run -it --rm --env BENCHDOG_HOST=h2load-server --env BENCHDOG_SCENARIOS=100:100 --env BENCHDOG_ITERATIONS=7 --env BENCHDOG_DURATION=60 --image quay.io/skupper/benchdog-h2load-client h2load-client