Synthetic monitoring: how to test your website automatically
Synthetic monitoring runs a planned test regularly, even when no customer is using your application. In DockRay, this test is called a scenario or flow. It describes a specific outcome, such as finding a product, changing a cart quantity or displaying payment methods.
Why is uptime monitoring not enough?
A server can respond successfully while the add-to-cart button does nothing. Checking the home page does not exercise a purchase journey. A scenario performs selected actions and compares the resulting state with your expectations. It complements uptime monitoring and error reporting; it does not replace your project's unit tests.
Two ways to test with DockRay
- Browser scenarios run a real Firefox browser through Camoufox. They support JavaScript, clicks, forms and waiting for elements. Choose this engine to test a store or application interface.
- Deep application checks call a public HTTPS endpoint implemented by your team. Your application executes the check and DockRay reads its result. This can verify a job queue or the completion of an import.
How do you set up a scenario?
- Open your project and its scenarios section. Choose the type of check.
- Name the test after the feature it verifies. Add browser steps and expectations, or enter the deep endpoint and its token.
- Set the interval, enable the scenario and run an initial test from the panel.
The effective frequency depends on the interval, your account plan and the engine's minimum interval. Browser runs have a separate budget. Check your panel for current limits. Manual runs also consume resources and remain subject to limits.
What does a test result mean?
A requested run first appears as queued, then running, and finally completed. A passing result means the defined conditions were met. A failure indicates an unsuccessful action or an unmet expectation. An inconclusive result means DockRay could not reliably assess the application, for example because it encountered a verification screen.
Failure notifications depend on enabled alerts, the consecutive-failure threshold and notification rate limits. One failed run does not necessarily trigger an immediate alert.
Does a scenario check the whole store?
It checks the journey you define. Create separate scenarios for separate processes, such as product search and the cart, so that a failure identifies the affected feature.
Continue with the shopping cart testing guide or learn about deep checks and post-deployment tests.