How resilience testing works
Resilience testing uses three core concepts:
-
Test template – A pre-configured, recommended test that defines which resilience capability to validate. A test template specifies the actions to perform, and declares the parameters it accepts.
-
Test – A test you create by configuring a recommended test template for your service, including your success criteria, stop conditions, and other settings. Each service has one test per template. For example, an Availability Zone: recovery test configured for your checkout service with a 30-minute RTO (Recovery Time Objective).
-
Test run – A single execution of a test. Each test run scopes to the current resources in your service and produces a pass or fail outcome, along with results and a report. You can have multiple runs of the same test. For example, running your Availability Zone: recovery test monthly to test recovery within your 30-minute RTO.
When you run a resilience test, the next generation of Resilience Hub performs the following:
-
Scopes the test – Automatically includes all resources discovered in your service that apply to the test.
-
Configures the test – Sets up the test with the configuration you provide, including success alarms, stop conditions, and log destinations.
-
Injects faults – Runs AWS FIS experiments to inject the faults defined by the test for the period of time specified.
-
Evaluates results – Determines a pass or fail outcome based on whether your service met your success criteria during the test. You can review test results in the console or programmatically. A test report is generated for each run.
Resilience tests are charged based on AWS FIS action-minute pricing. For details, see AWS Fault Injection Service pricing
