Unable to get flow variable inside exception. In below code I am trying to use centirofilename inside default-exception-strategy. It gives Exception
<set-variable value="#[xpath('//soap/filename/text()').text]"
variableName="centirofilename" doc:name="Variable" />
<default-exception-strategy>
<rollback-transaction exception-pattern="*" /> <!-- [1] -->
<processor-chain>
<logger level="INFO" category="ProTSP Logger"
message="#[centirofilename]" doc:name="Logger" />
</processor-chain>
</default-exception-strategy>
You should be using rollback-exception-strategy instead of the one you're using. The exception strategy you're using is legacy and it's use is not recommended.
Related
The question is simple. In some logback.xmls I see level as an element:
<logger name="mylog" additivity="false">
<level value="DEBUG" />
<appender-ref ref="fileAppender" />
</logger>
But in some it is written as an attribute:
<logger name="mylog" additivity="false" level="debug">
<appender-ref ref="fileAppender" />
</logger>
What is the difference?
Thanks.
In terms of configuring Logback, there is no difference. Both of the following declarations are functionally identical:
<logger name="com.x.y">
<level value="DEBUG"/>
</logger>
<logger name="com.x.y" level="DEBUG" />
Logback's configurer (have a look at ch.qos.logback.core.joran.GenericConfigurator.doConfigure()) creates an identical Logger instance for both of these declarations.
The only difference - when parsing the configuration - is that the first one manifests in more instances of ch.qos.logback.core.joran.event.SaxEvent (start and end events for the logger and for the level) than the second one (start and end events for the logger only).
If you are associating a logger with a specific appender then you'll already be defining a logger element body e.g.
<logger name="com.x.y">
<appender-ref ref="STDOUT"/>
</logger>
In which case defining a level within the element body rather than as an attribute might read better but it really is just a case of developer preference.
I am caching certain values in my azure api manager policy and in certain cases remove the item to clean up the cache and retrieve the value back from the api.
Based on my experience, even after I remove the value using the cache-remove-value policy, my next api call still finds the value in the cache. Here is a sample code:
<cache-store-value key="Key123" value="123" duration="300" />
<cache-lookup-value key="Key123" variable-name="CacheVariable" />
<cache-remove-value key="Key123" />
<cache-lookup-value key="Key123" default-value="empty" variable-name="CacheVariable2" />
<return-response>
<set-status code="504" reason="" />
<set-body>#(context.Variables.GetValueOrDefault<string>("CacheVariable2"))</set-body>
</return-response>
This code basically returns empty or "123" in the body based on if the cache item with key Key123 was found after being removed or not. This always returns the value of the cached item, "123".
Did anyone experienced this issue or found a way to clean up the cache?
If I continously check in a Retry, I can see that the item is sometimes cleaned after 2 seconds, sometimes 1 minute. I think the delete call is an async or queued call in the background so that we can't really be sure if it's cleaned or not without continuously checking.
UPDATE:
As an actual solution for now, instead of deleting, I actually update the cache item with 1 second duration and a dirty value.
This happens because cache removal request is asynchronous in regards to request processing pipeline, i.e. APIM does not wait for cache item to be removed before continuing with request, thus it is possible to still retrieve it right after removal request since it has not been sent yet.
Updated based on your scenario: why don't you try something like this then:
<policies>
<inbound>
<base />
</inbound>
<backend>
<retry condition="#(context.Response.StatusCode == 200)" count="10" interval="1">
<choose>
<when condition="#(context.Variables.GetValueOrDefault("calledOnce", false))">
<send-request mode="new" response-variable-name="response">
<set-url>https://EXTERNAL-SERVICE-URL</set-url>
<set-method>GET</set-method>
</send-request>
<cache-store-value key="externalResponse" value="EXPRESSION-TO-EXTRACT-DATA" duration="300" />
<!--... or even store whole response ...-->
<cache-store-value key="externalResponse" value="#((IResponse)context.Variables["response"])" duration="300" />
</when>
<otherwise>
<cache-lookup-value key="externalResponse" variable-name="externalResponse" />
<choose>
<when condition="#(context.Variables.ContainsKey("externalResponse"))">
<!-- Do something with cached data -->
</when>
<otherwise>
<!-- Call extenal service and store in cache again -->
</otherwise>
</choose>
<set-variable name="calledOnce" value="#(true)" />
</otherwise>
</choose>
<forward-request />
</retry>
</backend>
<outbound>
<base />
</outbound>
I've a polling process in Mule that queries a MySQL database every 30 seconds and sends an email to a recipient. How do I limit to sending just 1 email regardless of the polling cycle whether it be 30 seconds or 15 seconds? I'm open to a counter in the mysql db as well if that's an option.
Thank you.
Write a condition which will only send an email if emailSentFlag == false.
Use choice router to create the condition, and objectstore to hold the emailSentFlag value.
<flow...>
....
<objectstore:retrieve config-ref="ObjectStore__Configuration" key="emailSentFlag" defaultValue-ref="#[false]" targetProperty="flowVars.emailSentFlag" doc:name="retrieve emailSentFlag"/>
<choice doc:name="IsEmailSent?">
<when expression="#[flowVars.emailSentFlag == true]">
<logger level="INFO" doc:name="Log Email Already Sent"/>
</when>
<otherwise>
<smtp:outbound-endpoint host="" user="" password="" to="" from="" subject="test" cc="" bcc="" replyTo="" responseTimeout="10000" ref="Gmail" doc:name="SMTP" connector-ref="Gmail"/>
<objectstore:store config-ref="ObjectStore__Configuration" key="emailSentFlag" value-ref="#[true]" doc:name="store emailSentFlag"/>
</otherwise>
</choice>
</flow>
Also explore the TTL and Persistence feature of objectstore, it could be useful to you.
Cheers
You could use queueing(JMS) for your use case, before sending the "data" to jms would add a delay. You could do this like this
<message-properties-transformer overwrite="true" doc:name="Add DELAY in sending the Response to Queue">
<add-message-property key="Content_Type" value="application/json"/>
<add-message-property key="AMQ_SCHEDULED_DELAY" value="${aupost.retry.timeout}"/>
</message-properties-transformer>
Then add jms consumer to consume the message and send the email accordingly.
Do you have any sample flow to show? You may be able to use collection/message aggregators but looking at flow first would help to suggest.
You can put a vm queue in the end of your flow where the poller picks the data from the sql database.
In other flow, invoke the vm queue using a mule requester as an inbound connector inside a poll then set whatever frequency you want for the mail in the poll frequency using a cron expression or fixed-scheduler. Something like the below code:-
<flow name="db_poll">
<poll doc:name="Poll">
<db:no-operation-selected config-ref="" doc:name="Database"/>
</poll>
<logger message="invoking the database in the poll.. every 30 secs" level="INFO" doc:name="Logger"/>
<vm:outbound-endpoint exchange-pattern="one-way" path="email_queue" connector-ref="VMformail" doc:name="VM"/>
</flow>
<flow name="email_poll">
<poll doc:name="Poll">
<fixed-frequency-scheduler frequency="1" timeUnit="DAYS"/>
<mulerequester:request-collection resource="VMformail" timeout="100000" doc:name="Mule Requester"/>
</poll>
<logger message="send an email" level="INFO" doc:name="Logger"/>
<smtp:outbound-endpoint host="localhost" responseTimeout="10000" doc:name="SMTP"/>
</flow>
Trying this with logback suggests you can't set a level for an entire hierarchy. In other words, you can't specify something like:
<logger name="com.company.app.module.**" level="ERROR"/>
but instead you must specify:
<logger name="com.company.app.module.a" level="ERROR"/>
<logger name="com.company.app.module.a.b" level="ERROR"/>
<logger name="com.company.app.module.a.b.c" level="ERROR"/>
Is there no shorthand for an entire subpackage hierarchy?
I suggest that you read the manual and have a look at the example configuration. You specify a "hierarchy" without wildcard characters. Example
<logger name="com.company.app.module" level="ERROR"/>
<logger name="com.company.app.module.a" level="DEBUG"/>
<logger name="com.company.app.module.a.b" level="INFO"/>
The most specific logger wins. Effective level for com.company.app.module.a.b and below will be INFO. For com.company.app.module.a and below it will be DEBUG, except for com.company.app.module.a.b. And so on...
I use Mule Studio.
When I run for example localhost:8080/?first=value1&second=value2 I would like to get two variables and their values:
first: value1
second: value2
I use splitter to delete first '/' like this:
[regex('/(.*?)', message.payload)]
but now I get:
?first=value1&second=value2
You can extract the parameters by using message.inboundProperties['parameter'].
For example:
<logger level="WARN" message="#[message.inboundProperties['first']]" />
<logger level="WARN" message="#[message.inboundProperties['second']]" />
You may extract the parameters in three ways:
Directly from the message's inbound properties
As a Map by accessing the inbound property keyed with http.query.params
Using http:body-to-parameter-map-transformer and have the map conveniently placed on the payload.
Consider running the following flow:
<flow name="mule-configFlow1" doc:name="mule-configFlow1">
<http:inbound-endpoint address="http://localhost:8082/app" />
<http:body-to-parameter-map-transformer />
<logger level="ERROR" />
<logger level="ERROR" message="Payload is: #[payload]" />
<json:object-to-json-transformer />
</flow>