Goal: Distributed performance testing with Locust, Kubernetes, Elasticsearch.

Testing locally with Docker Compose

# build the testing server. simple nginx http server that return "Hello, World!"
docker build ./nginx -t test-server

# build the test locust container. it has a single test that hits the test-server
docker build ./locust -t my-locust

# start the components
# elasticsearch will keep its data in the "elasticsearch/data" directory
docker-compose up

• The test http server can be reached at http://localhost:8080/.
• The Locust web ui can be reached at http://localhost:3000/.
• Elasticsearch is avaiable at http://localhost:9200/.
• Kibana is available at http://localhost:5601/.

Create the index where the performance data will be stored. One way to do this is to use the Kibana console. Whichever way you choose, send a PUT request to the desired index. In this example that will be /performance_data. The example below is a mapping definition that will be sent in the performance_data index.

{
  "mappings": {
    "properties": {
      "request_type": { "type": "keyword" },
      "name": {
        "type": "text",
        "fields": {
          "raw": { "type": "keyword" }
        }
      },
      "response_time": { "type": "double" },
      "response_length": { "type": "integer" },
      "exception_message": {
        "type": "text",
        "fields": {
          "raw": { "type": "keyword" }
        }
      },
      "exception_type": { "type": "keyword" },
      "is_success": { "type": "boolean" },
      "timestamp": { "type": "date" },
      "host": {
        "type": "text",
        "fields": {
          "raw": { "type": "keyword" }
        }
      }
    }
  }
}