Testing and QA Environments

The key decision for your QA environment is driven by if you wish to conduct performance testing. If so, how much it should match the production environment in terms of scale. Some guidelines:

• For end-user performance testing of ‘search’ related transactions:

• The elastic cluster and web server should be scaled as production.

• The QA system should contain the same volume of content.

• If you have not scaled the QA platform to match production, you can extrapolate some estimations for performance. Please note this is not an exact science, and the only true test is to have the QA platform sized the same as production.

For strict pre-production testing environments:

• If production has a separate web server and agent servers ,which it should, then so should QA.

• If production has multiple agent servers, then QA should have at least 2 agent servers.

• If production has separate proxy nodes on the elastic cluster, then so should QA.

• If production uses a multi node elastic cluster than QA should do the same.

For performance testing Aiimi can provide separate guidelines for this. The most significant process for end user response times is search and aggregation requests.So, for performance testing the Elasticsearch cluster, and the web servers being scaled. If you want to test ingestion throughput ensure the Elasticsearch cluster and Agent servers are scaled.