Tracer Interceptor
Camel supports a tracer interceptor that is used for logging the route executions at INFO
level.
...
Implementing custom tracing via TracingInterceptor has been deprecated. However turning tracing on via setTracing(true) is okay.
Camel's tracer interceptor can be used for logging, by default at the INFO
level, a route's execution. The tracer is an instance of org.apache.camel.spi.InterceptStrategy
. It can be applied to either a DefaultCamelContext
or a The Tracer is an InterceptStrategy which can be applied to a DefaultCamelContext or SpringCamelContext
to ensure that there is a a TracerInterceptor
created for every node in the DSL.You can enable or disable the Tracerroute. The tracer's logging can be enabled/disabled dynamically , by calling the tracer's its setEnabled
method.
...
The org.apache.camel.processor.interceptor.TraceFormatter
has been rename to org.apache.camel.processor.interceptor.DefaultTraceFormatter
in Camel 2.0.
Users of Camel 1.x should use TraceFormatter
whereas 2.x users should use DefaultTraceFormatter
.
...
Checkout which dependencies are required by Camel for logging purpose.
...
From Camel 2.12: the CamelContext must be explicitly configured for tracing:
- XML:
<camelContext trace="true" ...>
- Java:
camelContext.setTracing(true)
.
Options
...
Options
Trace has been improved in Camel 1.5 to be more configurable with these options:
...
Option
...
Default
...
Description
...
formatter
...
...
Option | Default | Description |
---|---|---|
|
| Optional destination URI to route TraceEventExchange containing TraceEventMessage with details about the trace. Can be used for custom processing to store traces in database using JPA. |
|
| Enable/disable the tracer. |
|
| Sets the Trace Formatter to use. Default: |
...
|
...
|
...
true
...
Flag to enable or disable this tracer
...
logLevel
...
INFO
...
The logging level to use: FATAL, ERROR, WARN, INFO, DEBUG, TRACE
...
logName
...
...
| Camel 2.3: Fully qualified class name for a custom |
...
|
...
traceFilter
...
null
...
An exchange Predicate to filter the tracing.
...
traceInterceptors
...
false
...
Flag to enable or disable tracing of interceptors
...
traceExceptions
...
true
...
Flag to enable or disable tracing of thrown exception during processing of the exchange
...
traceOutExchanges
...
false
...
Flag to enable fine grained tracing with a callback for both IN (before) and OUT (after). Is disabled by default which means there is only one trace callback executed.
...
logStackTrace
...
false
...
Camel 2.0: When tracing exception you can control whether the stack trace should be logged also. If not then only the exception class and message is logged.
...
useJpa
...
false
...
Camel 2.0: To use a JpaTraceEventMessage
from camel-jpa component as the TraceEventMessage
. This requires that camel-jpa.jar is on the classpath.
...
destinationUri
...
null
...
Camel 2.0: Optional destination uri to route TraceEventExchange containing TraceEventMessage with details about the trace. Can be used for custom processing to store traces in database using JPA.
...
For example you can use your custom JPA @Entity class to store traced information in a database according to your schema layout. | ||
|
| The logging level to use: |
|
| The log name to use. Default: |
|
| Controls whether the stack trace of a traced exception should be logged. If |
|
| An Exchange Predicate to filter the tracing. |
|
| Enable/disable tracing of interceptors. |
|
| Enable/disable tracing of an exception thrown during the processing of an Exchange. |
|
...
| Camel 2.3: |
...
To use a custom |
...
|
...
traceHandler
...
null
...
| ||
|
| Camel 2.3: To use a custom |
...
Default: |
Formatting
The tracer formats the execution of exchanges to log lines. They are logged at INFO level in the log category: org.apache.camel.processor.interceptor.TraceInterceptor
.
The tracer uses by default TraceFormatter
to format the log line.
TraceFormatter has the following options:
...
Option
...
Default
...
Description
...
breadCrumbLength
...
0
...
Camel 1.5.1: Fixed length of the bread crumb. 0 = no fixed length. Setting a value to e.g. 80 allows the tracer logs to be aligned for easier reading.
...
nodeLength
...
0
...
|
| Enable/disable fine grained tracing, using a callback, for both By default only one trace callback is executed. |
|
| To use a |
Formatting
The tracer formats the execution of exchanges to log lines. They are logged at INFO
level in the log category: org.apache.camel.processor.interceptor.TraceInterceptor
. By default the tracer uses: org.apache.camel.processor.interceptor.DefaultTraceFormatter
to format the log line.
DefaultTraceFormatter
has the following options:
...
Option | Default | Description |
---|---|---|
|
| Fixed length of the bread crumb.
Setting a value to e.g. |
...
|
...
|
...
|
...
Limits the |
...
number of chars logged per line. From Camel 2.9: the default value is | ||
|
| Camel 2.18: If |
|
| Fixed length of the node.
Setting a value to e.g. |
|
| Output the |
|
| Output the |
|
| Outputs the unique UnitOfWork for the exchange. Can be used for correlation to identify a particular Exchange. |
|
| Output the exception if the processing of an Exchange has failed. |
|
| Enable/disable the output of an Exchange's unique Currently the breadcrumb is sufficient. |
|
| Output the Message Exchange Pattern (MEP). |
|
| Output the |
|
| Previous and destination node. Displayed as: |
|
| Output the |
|
| Output the |
|
| Output the |
|
| Output the Exchange's properties. |
|
| Camel 2.8: output the |
|
| To output the Exchange's unique |
...
From Camel 2.8: the Camel Tracer will by default not log stream or files bodies. To force Camel to log these set the following property on the CamelContext properties:
...
Example:
...
where:
ID-claus-acer/3690-1214458315718/2-0
is the breadcrumb with the unique correlation id.-
node3
is the id of the node in the route path. Always displayed. -
To[mock:a]
is the destination node. -
InOnly
is the exchange pattern. Is always shown. - Then the rest is properties, headers and the body.
Showing from
and to
The trace log will output both the from and to so you can see where the Exchange came from, such as:
...
Enabling
To enable tracer from the main run:
...
or
...
and the tracer will be activated.
Enabling in Java
...
You can configure tracing
...
showNode
...
true
...
In Camel 1.x its the destination node. In Camel 2.0 its both the previous and destination node, so you can see from -> to.
...
showExchangeId
...
false
...
To output the unique exchange id. Currently the breadcrumb is sufficient.
...
showShortExchangeId
...
false
...
Camel 1.5.1: To output the unique exchange id in short form, without the hostname.
...
showProperties
...
true
...
Output the exchange properties
...
showHeaders
...
true
...
Output the in message headers
...
showBodyType
...
true
...
Output the in body Java type
...
showBody
...
true
...
Output the in body
...
showOutHeaders
...
false
...
Camel 2.0: Output the out (if any) message headers
...
showOutBodyType
...
false
...
Camel 2.0: Output the out (if any) body Java type
...
showOutBody
...
false
...
Camel 2.0: Output the out (if any) body
...
showExchangePattern
...
true
...
Camel 1.5: Output the exchange pattern
...
showException
...
true
...
Camel 1.5: Output the exception if the exchange has failed
...
maxChars
...
...
Camel 2.0: Is used to limit the number of chars logged per line.
Example:
Code Block |
---|
ID-claus-acer/4412-1222625653890/2-0 -> to(mock:a) , Pattern:InOnly , Headers:{to=James} , BodyType:String , Body:Hello London
|
Wiki Markup |
---|
{{ID-claus-acer/3690-1214458315718/2-0}} is the breadcrumb with the unique correlation id.
{{node3}} is the id of the node in the route path. Is always shown.
{{To\[mock:a\]}} is the destination node.
{{InOnly}} is the exchange pattern. Is always shown.
Then the rest is properties, headers and the body. |
Showing from and to
In Camel 2.0 the trace log will output both the from and to so you can see where the Exchange came from, such as:
Code Block |
---|
>>> direct:start --> process(MyProcessor)
>>> process(MyProcessor) --> to(mock:a)
>>> to(mock:a) --> to(mock:b)
|
Enabling
To enable tracer from the main run
Code Block |
---|
java org.apache.camel.spring.Main -t
|
or
Code Block |
---|
java org.apache.camel.spring.Main -trace
|
and the tracer will be active.
Enabling from Java DSL
The tracer can be enabled by adding it to the interceptor chain to the camel context. This is demonstrated in the unit test below.
Notice: We could have changed the properties on the Tracer object before adding it, if we e.g. don't like the default settings.
Wiki Markup |
---|
{snippet:id=e1|lang=java|url=camel/branches/camel-1.x/camel-core/src/test/java/org/apache/camel/processor/TraceInterceptorTest.java} |
In Camel 2.0 this is a bit easier as you just do:
Code Block |
---|
context.setTracing(true);
|
In Camel 2.0 you can also configure it at a higher granularity as you can configure it on camel context and then override and set it per route as well. For instance you could just enable tracer for one particular route.
Running the test we get the trace information logged at INFO level (Camel 1.x output):
Code Block |
---|
INFO TraceInterceptor - ID-claus-acer/4412-1222625653890/2-0 -> process(MyProcessor) , Pattern:InOnly , Headers:{to=James} , BodyType:String , Body:Hello London
INFO TraceInterceptor - ID-claus-acer/4412-1222625653890/2-0 -> to(mock:a) , Pattern:InOnly , Headers:{to=James} , BodyType:String , Body:Hello London
INFO TraceInterceptor - ID-claus-acer/4412-1222625653890/2-0 -> to(mock:b) , Pattern:InOnly , Headers:{to=James} , BodyType:String , Body:Hello London
...
INFO TraceInterceptor - ID-claus-acer/4412-1222625653890/2-1 -> process(MyProcessor) , Pattern:InOnly , Headers:{from=Claus} , BodyType:String , Body:This is Copenhagen calling
INFO TraceInterceptor - ID-claus-acer/4412-1222625653890/2-1 -> to(mock:a) , Pattern:InOnly , Headers:{from=Claus} , BodyType:String , Body:This is Copenhagen calling
INFO TraceInterceptor - ID-claus-acer/4412-1222625653890/2-1 -> to(mock:b) , Pattern:InOnly , Headers:{from=Claus} , BodyType:String , Body:This is Copenhagen calling
|
And the same output with Camel 2.0:
...
well. For instance you could just enable the tracer for a particular route.
...
Configuring
...
in Java
...
The tracer Tracer options can be configured from the Java DSL like this:
...
in Java as follows:
...
Using
...
Predicates to
...
Filter Exchanges
Available as of Camel 1.5
In the code below we want the tracer only to trace if the body contains the text London
. As this is just an example can of course set any Predicate that matches your criteria:
...
Enabling
...
in Spring XML
There is now a a trace
attribute you can specify on the *the <camelContext/>
for example.
...
Example:
...
You can see this in action with the SpringTraceTest and its spring.xml file
Another option is to just include a spring XML which defines the Tracer bean such as the one that is automatically included if you run the Main with -t above.
...
Configuring in Spring
...
XML
You In Camel 1.5 you can configure the tracer as a Spring bean. Just add a bean with the bean class org.apache.camel.processor.interceptor.Tracer
and Camel will use it as the Tracer.
...
as the tracer.
...
You
Formatting from Spring
In Camel 1.5 you can configure the formatting of tracer as a Spring bean. Just add a bean with the the id="traceFormatter"
and Camel will lookup this this id
and use the formatter, as the example below illustrates:
...
associated formatter.
Example:
...
Enable Tracing of OUT
Messages
To trace the
Enable tracing of out messages
Available as of Camel 2.0
You can now trace messages coming out of processing steps . To enable this, configure the tracer as follows
...
:
...
or
...
...
Running with With these options , you'll get output similar to:
...
the output will look like:
...
Using a Custom Formatter
Avaiable as of Camel 2.0
You can now implement your own To create a custom formatter create a class that implements the interface org.apache.camel.processor.interceptor.TraceFormatter
to be used for logging trace messages to the log.
The sample below shows how to configure a Tracer from Java DSL using custom formatter:
...
.
Example:
And here we have our custom logger that implements the TraceFormatter
interface where we can construct the log message how we like:
...
Using a Destination for
...
Custom Processing and Routing
Tracer now supports custom processing of trace events. This can be used to route a trace event to a JPA endpoint for persistence in a database.
This works by Camel creates a new TraceEventExchange TraceEventMessage containing:
- snapshot of the original traced Exchange as a immutable TraceEventMessage containing String values of the fields, when the interception occurred. This ensures the fields contains the exact data at the given time of interception.
- the original Exchange can in some implementations be accessed using
getTracedExchange()
...
- (though with JPA based tracer you cannot get the original Exchange).
...
Beware to access the original Exchange to avoid causing any side effects or alter its state. Prefer to access the information from TraceEventMessage
Camel routes the TraceEventExchange the TraceEventMessage
synchronously from the point of interception. When its completed Camel will continue routing the original Exchange.
The sample below demonstrates this feature, where we route traced Exchanges to the direct:traced
route:
...
Then we can configure a route for the traced messages:
...
And our processor where we can process the TraceEventMessage
. Here we want to create a CSV format of the trace event to be stored as a file. We do this by constructing the CSV String and the replace the the IN
body with our String instead of the TraceEventMessage
.
...
of the TraceEventMessage
.
Using JPA as
...
a Datastore for Trace Messages
Avaiable as of Camel 2.0
See Tracer Example for complete documentation and how to use this feature.
Traced
...
Route Path During Runtime
Available as of Camel 2.0
Tracer also traces the actual route path taken during runtime. Camel will store the route path taken on the UnitOfWork when Tracer is enabled. The example below demonstrates how we can use that for error handling where we can determine at which node in the route graph the error triggered.
First we define our route:
...
...
And then our custom error processor where we can handle the exception and figure out at which node the exception occurred.
...