Load testing is the process of putting demand on a system and measuring its response. You can conduct load testing:
• Proactively when planning for a new configuration to set a baseline for your account.
• When implementing a new functionality, such as during new releases.
• Validating account throughput before a planned peak season. When identifying your load test parameters, ensure you have specific and realistic performance and load targets.
You also want to avoid taking on too much and trying to test everything; load testing should always be done as
targeted approach.
Load tests should have a specific and realistic performance and load target.
Performance targets can be extrapolated from an individual action’s execution time.
For example, if a Sales Order Save performs at 10 seconds per save, and you are testing 1,000 sales order saves, then you can reasonably expect that will take 10,000 seconds (about 2 hours and 45 minutes). available concurrency.
Concurrent threads may not increase throughput linearly.
Concurrent access to resources may affect performance.
You should also ensure your test data represents live data and use patterns. Test data should account for the normal
variation in your expected production data.
Additionally, you should not use a single customer or item when testing. For example, if you are testing the process
of a customer paying for an order, consider the following best practices:
• The customer should vary.
• The payment type should vary.
• The number, type, and quantity of order lines
should vary.
• Test order and order line customization
Note that stress testing is different than load testing. Stress
testing is a form of deliberately intense testing used to
determine the stability of a given system. It involves testing
beyond normal operational capacity, often to a breaking
point, to observe the results. Stress testing is not supported
in NetSuite.