I am using MULE and 3 transformers, 2 that that require information from an HTTP response + a POJO. Transformer 2 and 3 only get the HTTP POST repsonse now and not the POJO.
Transformer1 gets a POJO and transforms it , sends it to a HTTP endpoint, it uses a Post request.
The response of the POST is used in transformer 2 and 3, however they
also need the POJO that to do their transformation !
How can I use the pojo that was used in Transformer 1 + the http response in transformer
2 and 3? I want a message to enter the ALL control flow with the payload = http response + the POJO.
Is it possible to add the two in a message e.g 2 payloads in 1 message ? How can I solve this issue?
Thanks in advance.
Mule Flow:
You can store the original POJO payload as a message property to access it later in the flow. See here for an introduction to the Mule message property scopes:
http://blogs.mulesoft.org/mule-school-the-mulemessage-property-scopes-and-variables/
Related
With Mule 3 it was possible to send messages asynchronously to an endpoint using MuleClient:
MuleClient client = new MuleClient(muleContext);
client.dispatch("vm://vm.queue", "Message Payload", null);
Is there a way to migrate this functionality in Mule 4 since MuleClient has been removed?
I came across a post that suggested getting the flow by name and publishing the message to the flow as follows
Flow flow = registry.lookupByName("MyFlow").get();
InputEvent event = new DefaultInputEvent();
event.message(Message.of(payload));
flow.execute(event);
but I get a ClassNotFoundException for the class org.mule.runtime.internal.event.DefaultInputEvent
Using Harshank's recommendation I was able to push messages to a flow simply by getting a reference to the flow and triggering the flow by sending messages to the source.
Flow flow = registry.lookupByName(flowName).get();
ComponentLocation location = DefaultComponentLocation.from(flowName + "/source");
...
Message message = Message.of(payload);
CoreEvent coreEvent = CoreEvent.builder(EventContextFactory.create(flow, location)).message(message).build();
flow.process(coreEvent);
This is a much cleaner solution than what is implemented in the blog and works from beans initialized in the Spring module. As aled mentioned, this is bad practice, but in the interest of time it is a solution.
We are migrating from Mule 3 to Mule 4 and in one of our functionalities we need to publish messages to a topic and downstream another mule component is consuming from the queue which is bridged to the topic.
Nothing special here .
To ensure we are able to trace the flow via logs we were sending a 'TrackingId' attribute while publishing messages to the topic ( Mule 3 )
message.setOutboundProperty("XYZ_TrackingID", flowVars['idFromUI']);
return payload;
However when I try the same in Mule 4 we get the following exception :
ERROR 2020-12-20 10:09:12,214 [[MuleRuntime].cpuIntensive.14: [mycomponent].my_Flow.CPU_INTENSIVE
#66024695] org.mule.runtime.core.internal.exception.OnErrorPropagateHandler:
Message : groovy.lang.MissingMethodException: No signature of method:
org.mule.runtime.api.el.BindingContextUtils$MessageWrapper.setOutboundProperty() is applicable for
argument types: (java.lang.String, org.mule.weave.v2.el.ByteArrayBasedCursorStream) values:
[XYZ_TrackingID, "1234567"].\nError type : (set debug level logging or '-
Dmule.verbose.exceptions=true' for
everything)\n********************************************************************************
Checked internet and it seems in Mule4 setting outbound properties is removed as per here
So how do I achieve the same in Mule 4 ?
Don't even try to do that for several reasons. For one message structure is different, so output properties doesn't exist anymore and that method doesn't even exists. On the other hand, in Mule 4 components like the Groovy component can only return a value and cannot change the event. They can not decide to what that value is going to be assigned. You can set the target in the configuration (payload or a variable) and not change the attributes. Note that variables in Mule 4 are referenced by var., not by flowVars. like in Mule 3 (ie vars.idFromUI).
There is a simpler way to set message properties in the Mule 4 JMS connector. Use the properties element and pass it an object with the properties.
For example it could be something like this:
<jms:publish config-ref="JMS_config" destination="${bridgeDestination}" destinationType="TOPIC">
<jms:message>
<jms:body>#["bridged_" ++ payload]</jms:body>
<jms:properties>#[{
XYZ_TrackingID: vars.idFromUI
}]</jms:properties>
</jms:message>
</jms:publish>
It is in the documentation: https://docs.mulesoft.com/jms-connector/1.0/jms-publish#setting-user-properties. I adapted my example from there.
I am not sure if Correlation Id serves the purpose of a tracking ID for your scenario. But you can pass a CID as below. It's there in the mule documentation.
https://docs.mulesoft.com/jms-connector/1.7/jms-publish
<jms:publish config-ref="JMS_config" sendCorrelationId="ALWAYS" destination="#[attributes.headers.replyTo.destination]">
<jms:message correlationId="#[attributes.headers.correlationId]"/>
</jms:publish>
If your priority is to customise the Tracking ID you want to publish, then try passing below format. The key names may differ as per your use case.
<jms:publish config-ref="JMS_config" destination="${bridgeDestination}" destinationType="TOPIC">
<jms:message>
<jms:body>#["bridged_" ++ payload]</jms:body>
<jms:properties>#[{
AUTH_TYPE: 'jwt',
AUTH_TOKEN: attributes.queryParams.token
}]</jms:properties>
</jms:message>
</jms:publish>
In the above the expression attributes.queryParams.token is basically trying to access a token query parameters which is passed to JMS as a property AUTH_TOKEN key-name , consumed by the API through a HTTP Listener or Requestor earlier.
However, attributes.headers.correlationId is a header. Both queryParams and headers are part of attributes in Mule 4.
my requirement is to sent the error http status and error message in the body.
In Case of error in flow i need to pass the http code in the status field.
I can configure this http listner but don't know how to set this to get into payload. Please guide on that.
I'm expecting the MEL to get the 400 Bad Request
{
'status': "400 Bad Request",
'message': error.description
}
I could think of 2 ways of doing this:
1.You can use RAML and for every status code you can send appropriate responses as per your use case.This I think is the best way of doing it.
2.You can have the value for status and message key of your response body inside error handling block.Configure the 'TYPE' condition of your error handling block to catch a certain HTTP error message , then inside that you can set 2 vars: one with the status value and other with message value.Then use this vars in the Error response section of HTTP listener. You'll be ending up with many such error handling blocks if you want to address multiple status codes.
Let me know if you need further clarifications.Hope it helps.
Say for example, I created a flow for scatter-gather and I want to check if all endpoints are returning same result status code 200 or throw an error if not.
Configure the Response Validator (General > Response > Response Validator) for each HTTP Request so only 200..299 responses are considered valid.
You can use try block for every HTTP request on wrap whole scatter gather. If one fails, capture HTTP status code in on error propogate and log the results.
I suggest you wrap each request into try block, if you already have a global error handler defined, it should pick up status code 500 etc. Otherwise, capture response code into dataweave
Is there any way to use the response of a wcf service method request as an input for the next request in same orchestration and return the response of the first request as well as the response of the second request as out put in BizTalk?
Eg :
My first request gives a response as "a"
Give this response "a" as request to the 2nd request and gets the response as "b"
Return the response as "a" and "b".
Is this possible?
Yes. You could either create a map from Response 1 to Request 2, and also create a multiple input message map from Response 1 and Response 2 to your final output message.
If the messages involved don't have any repeating structures, it may be enough to distinguish the fields that you need to be concerned with and just use a ConstructMessage with an XmlDocument, i.e.
// construct shape code
varXmlDoc = new System.Xml.XmlDocument();
varXmlDoc.LoadXml("<webSvcRequest2 xmlns=''><ParamB>" + msgWebSvcResp1.ParamA + "</ParamB></webSvcRequest2>");
msgWebSvcReq2 = varXmlDoc;
And similar code for producing the final output message. If you go this route, I'd advice creating some C# utility methods to actually store the strings/message templates.