How To:
Most workloads are configured to send the benchmark data to Elasticsearch if given the proper information.
To enable this functionality a few variables must be set in the workload CR file.
elasticsearch:
server: the elasticsearch server to upload to
port: the elasticsearch server port
parallel: enable parallel uploads to elasticsearch [default: false]
index_name: the index name to use [default: workload defined]
NOTE: The only required parameters if using elasticsearch and the port and server. The others are optional and will be defaulted if not provided. Additionally, enabling parallel uploading may greatly impact Elasticsearch server performance. Ensure your environment is configured to handle the increased load before enabling
Example CR with elasticsearch information provided
apiVersion: ripsaw.cloudbulldozer.io/v1alpha1
kind: Benchmark
metadata:
name: smallfile-benchmark
namespace: my-ripsaw
spec:
test_user: homer_simpson
clustername: test_ci
elasticsearch:
server: my.es.server
port: 8080
index_name: ripsaw-smallfile
metadata:
collection: true
workload:
name: smallfile
args:
clients: 2
operation: ["create", "read", "append", "delete"]
threads: 1
file_size: 0
files: 100000