Introduction › Service Concepts › Propagation Types › Custom Propagation

Custom Propagation

Custom propagation indicates that one CI may depend on several other CIs for some behavior or function. Custom propagation lets you specify policy that defines when and how to change the severity value of a parent item in a dependent relationship. If the policy is not met, there is no impact to the CI or service.

Usage

Use Custom propagation in the following situations:

• When a group of CIs work together to deliver an aspect of the service
• When a parent CI is not directly affected by a fault condition in an individual CI and is instead affected by the combined performance of all child CIs

Examples

Examples of appropriate situations to use custom propagation include the following:

• A Group CI for a web server farm may require custom propagation with its servers to indicate that the farm can lose 30% of its servers and still function.
• A clustered server configuration with failover capacity may require custom propagation to indicate that a certain number of failures are permitted as long as the failover servers are functioning.

USM Relationships

The following USM relationships map to custom propagation by default in CA SOI:

• Is Affected By
• Is Evolution Of

You must select one of these relationships to effect the custom propagation behavior. The relationships define that the impact from one CI to another is derived through configurable policy.

Custom propagation requires you to create a propagation policy that defines how to calculate the impact from a group of related CIs to a parent CI. You can configure CA SOI to automatically assign custom propagation and the associated propagation policy when a CI is added to a group that already uses custom propagation instead of the default aggregate propagation.

When you assign custom propagation policy to a CI and the threshold is reached, the target CI attains the specified severity value and an infrastructure alert with the appropriate severity. However, because the alert applies to the CI as it relates to the parent service, the CI icon does not change color as expected, and the severity propagates to a separate impact calculation on the service, which is reflected in the service icon color.