Testing and Mocking Errors

Expect an Error in your Test

During the execution of a flow, components may throw errors which might be expected for the use case being tested. In these cases, you want to specify that your test expects a particular error type or exception.

For example, lets look at the following Mule code:

<flow name="retrieveKey">
    <os:retrieve key="#[payload.key]"/>
    <set-variable value="#[payload.id]" variableName="userId"/>
</flow>

The key is taken from the payload and if the key is null, the Object Store connector will throw an error with OS:INVALID_KEY error type.

The test to validate this case would be something like this:

<munit:test name="retrieveNullKeyTest" description="Expect error when key is null" expectedErrorType="OS:INVALID_KEY">
    <munit:behavior>
        <munit:set-event>
            <munit:payload value="#[ { key: null}]" />
        </munit:set-event>
    </munit:behavior>
    <munit:execution>
        <flow-ref name="retrieveKey"/>
    </munit:execution>
</munit:test>

Error Handlers Testing

In many cases, errors are handled by the user. They may want to perform some actions when an error is thrown, and decide to propagate the error or continue with the flow’s execution.

We can have different tests for each of the following possibilities:

On Error Continue

For example, lets look at the following mule code:

<flow name="retrieveKey">
    <os:retrieve key="#[payload.key]"/>
    <set-variable value="#[payload.id]" variableName="userId"/>
    <error-handler >
        <on-error-continue type="OS:INVALID_KEY">
            <set-payload value="#[{ message : 'Key is null'}]" />
        </on-error-continue>
    </error-handler>
</flow>

The OS:INVALID_KEY error is being handled and the execution continues with a specific payload.

Since the error is not propagated, the execution resumes and the event can be verified:

<munit:test name="retrieveNullKeyTest" description="Expect message when key is null">
    <munit:behavior>
        <munit:set-event>
            <munit:payload value="#[ { key: null}]" />
        </munit:set-event>
    </munit:behavior>
    <munit:execution>
        <flow-ref name="retrieveKey"/>
    </munit:execution>
    <munit:validation >
        <munit-tools:assert-that expression="#[payload.message]" is="#[MunitTools::equalTo('Key is null')]"/>
    </munit:validation>
</munit:test>

As you can see, in this case you don’t have to expect the OS:INVALID_KEY error type.

On Error Propagate

For example, if the user’s code is the following:

<flow name="retrieveKey">
    <os:retrieve key="#[payload.key]"/>
    <set-variable value="#[payload.id]" variableName="userId"/>
    <error-handler >
        <on-error-propagate type="OS:INVALID_KEY">
            <raise-error type="APP:NULL_KEY"/>
        </on-error-propagate>
    </error-handler>
</flow>

The OS:INVALID_KEY error is being handled and the error APP:NULL_KEY is raised.

Since the test will fail with the raised custom error type, the test should expect it:

<munit:test name="retrieveNullKeyTest" description="Expect APP:NULL_KEY is raised" expectedErrorType="APP:NULL_KEY">
    <munit:behavior>
        <munit:set-event>
            <munit:payload value="#[ { key: null}]" />
        </munit:set-event>
    </munit:behavior>
    <munit:execution>
        <flow-ref name="retrieveKey"/>
    </munit:execution>
</munit:test>

Mock Errors

When you don’t want to trigger manually an error on a component (like the examples above), but you only want to make the processor fail with a specific error, you can use the error parameter of the mock-when processor.

For example, if the user’s code is the following:

<flow name="requestFlow">
    <http:request method="GET" config-ref="HTTP_Request_configuration"/>
    <set-variable value="#[payload.id]" variableName="userId"/>
    <error-handler >
        <on-error-continue type="HTTP:CONNECTIVITY">
            <logger level="ERROR" message="AN ERROR OCCURRED"/>
        </on-error-continue>
    </error-handler>
</flow>

and you want to make the http:request fail with an HTTP:CONNECTIVITY error and also check that the logger was actually executed, the test may look something like this:

Mock Error with On Error Continue

<munit:test name="requestFlowTest" description="Mock CONNECTIVITY error" >
    <munit:behavior >
        <munit-tools:mock-when processor="http:request">
            <munit-tools:then-return >
                <munit-tools:error typeId="HTTP:CONNECTIVITY" />
            </munit-tools:then-return>
        </munit-tools:mock-when>
    </munit:behavior>
    <munit:execution >
        <flow-ref name="requestFlow"/>
    </munit:execution>
    <munit:validation >
        <munit-tools:verify-call processor="mule:logger"/>
    </munit:validation>
</munit:test>

If the On Error scope is an on-error-propagate, the verifications won’t be executed since the flow will propagate the error and fail, then the validation section of the test will be skipped.

In this case the error could be caught with a try scope and then the verifications can be performed:

Mock Error with On Error Propagate

<munit:test name="requestFlowTest" description="Mock CONNECTIVITY error" >
    <munit:behavior >
        <munit-tools:mock-when processor="http:request">
            <munit-tools:then-return >
                <munit-tools:error typeId="HTTP:CONNECTIVITY" />
            </munit-tools:then-return>
        </munit-tools:mock-when>
    </munit:behavior>
    <munit:execution >
        <try>
            <flow-ref name="requestFlow"/>
            <error-handler >
                <on-error-continue type="HTTP:CONNECTIVITY"/>
            </error-handler>
        </try>
    </munit:execution>
    <munit:validation >
        <munit-tools:verify-call processor="mule:logger"/>
    </munit:validation>
</munit:test>

Set an Event with an Error

When extracting error handling logic into a separated flow or sub flow, the input event must have an error that can be used in the error handling logic.

For example, lets look at the following mule code:

<flow name="retrieveFlow">
    <os:retrieve key="#[payload.key]"/>
    <error-handler >
        <on-error-continue>
            <flow-ref name="error-handlingSub_Flow"/>
        </on-error-continue>
    </error-handler>
</flow>

<sub-flow name="error-handlingSub_Flow">
    <logger level="INFO" message="#['Execution failed with: $(error.description)']"/>
</sub-flow>

If you wish to test the subFlow independently, you will need to set an error in the event.

The test may look something like this:

Set Error in the Event

<munit:test name="error_handlingSub_FlowTest" description="Test subFlow" >
    <munit:behavior >
        <munit:set-event>
            <munit:error id="OS:INVALID_KEY" exception="#[java!java::lang::Exception::new('Invalid Key')]"/>
        </munit:set-event>
    </munit:behavior>
    <munit:execution >
        <flow-ref name="error-handlingSub_Flow"/>
    </munit:execution>
    <munit:validation >
        <munit-tools:verify-call processor="mule:logger">
            <munit-tools:with-attributes >
                <munit-tools:with-attribute attributeName="message" whereValue="#['Execution failed with: Invalid Key']" />
            </munit-tools:with-attributes>
        </munit-tools:verify-call>
    </munit:validation>
</munit:test>