Summary

CloudStack currently generates many API error messages through ad-hoc string construction, leading to inconsistencies, limited customization, and no clear support for localization.
This feature introduces a structured, key-based error messaging framework where the management server emits a stable error message key along with contextual metadata. Clients and operators can use these to present consistent, customizable, and localized error messages.

Goals

Non-Goals

High-Level Design

Error Message Keys and Metadata

Errors are identified using a unique error message key (for example, vm.deploy.serviceoffering.inactive) and accompanied by a metadata map containing contextual information such as IDs, names, or related objects.

Instead of hard-coded strings, exceptions are raised with:

API Response Changes

API error responses include the following additional fields:

The existing errortext field is preserved for backward compatibility.

Example:

{
	"deployvirtualmachineresponse": {
		"errorcode": 431,
		"cserrorcode": 4350,
		"errortext": "Unable to deploy Instance as the given service offering 'test' (ID: 13, UUID: ...) is inactive. Specify an active service offering.",
		"errortextkey": "vm.deploy.serviceoffering.inactive",
		"errormetadata": {
			"serviceOffering": "'test' (ID: 13, UUID: ...)"
		}
	}
}

Error Message Templates

Template Source

Error message templates are stored in a JSON file on the management server:

/etc/cloudstack/management/error-messages.json

The default file is shipped with CloudStack and operators may modify it to customize messages.

Template Format

Each entry maps an error message key to a message template.

Error message keys follow a structured, hierarchical naming convention to ensure consistency, readability, and easy discoverability.

The general format is:

<actionable_resource>.<action>.<failing_resource_or_entity>.<cause_and_context>...

Where:


Templates support placeholder substitution using {{placeholder}} syntax.

Admin-specific variants are supported by adding a .admin suffix to the key.

Example:

{
	"vm.deploy.serviceoffering.inactive":
	"Unable to deploy Instance as the given service offering {{serviceOffering}} is inactive. Specify an active service offering.",
	"vm.deploy.serviceoffering.inactive.admin":
	"Unable to deploy Instance as service offering {{serviceOffering}} is inactive. Please activate it or choose a different offering."
}

ErrorMessageResolver

A new component, ErrorMessageResolver, is responsible for resolving error message keys into user-facing messages.

Responsibilities:

Role-Aware Error Messages

The caller’s role is determined using CallContext.
If the caller is a root admin and an admin-specific template exists, it is used; otherwise, the default template is applied.

Exception Handling Changes

Common exceptions (such as CloudRuntimeException and ServerApiException) are enhanced with constructors that accept:

error message key

metadata map

During API response generation:

The error message key and metadata are added to the response.

The resolved message is used to populate errortext.

Initial Scope

The initial implementation applies this framework to selected validation paths in the VM deployment flow.
Additional flows and modules can be incrementally migrated in future releases.

Backward Compatibility

Packaging and Operations

Localization

Localization of error messages is handled at the client side.

The CloudStack management server returns a stable error message key (errortextkey) along with contextual metadata (errormetadata). Clients are expected to resolve this key into a localized, user-facing message.

CloudStack UI Localization

When using the official CloudStack UI:

This approach:

Other API Clients

Other API clients (CLI tools, SDKs, external integrations) may:

Testing Strategy*

Unit tests for:

Integration tests validating API responses include error message keys and metadata.

Future Enhancements