Purpose:

Injects random failures, exceptions and delays.

Useful for chaos engineering, like promoted in DevOps practices, to test the resilience of the applications.

If an application has atomic transactions and handles well exceptions and timeouts, it should remain stable and coherent, given any exception or delay. 

It is also useful to test offline apps, to simulate random timeouts and errors at the server when syncing.

How to:

To use it, just drag and drop the action FailureInjection to all server side actions, at the least the top layer ones that are referenced by the interfaces, either UI or API's.

This implementation is just for server side.

Configurations:

By default all Site Properties are disabled to ensure this is promoted to production in a safe way.

By default the chance of an injection is 20%.

By default the max delay is of 15s. The actual delay is random between 0-maxdelay.

The action is protected by two site properties that need to be enabled to avoid unaware activation in production environments, EnableFailureInjGeneral and EnableFailureInjDoubleSecurity.

If this is enabled, a message will appear in the error and general log like follows:

"FAILURE INJECTION - DISABLE AT MODULE FailureInjection - SET SITE PROPERTY EnableFailureInjGeneral and EnableFailureInjDoubleSecurity TO FALSE"