org.apache.camel.processor.interceptor.DefaultChannel Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of camel-core Show documentation
Show all versions of camel-core Show documentation
The Core Camel Java DSL based router
/**
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.camel.processor.interceptor;
import java.util.ArrayList;
import java.util.Collections;
import java.util.List;
import java.util.Map;
import org.apache.camel.AsyncProcessor;
import org.apache.camel.CamelContext;
import org.apache.camel.CamelContextAware;
import org.apache.camel.Channel;
import org.apache.camel.Exchange;
import org.apache.camel.Processor;
import org.apache.camel.management.InstrumentationInterceptStrategy;
import org.apache.camel.management.InstrumentationProcessor;
import org.apache.camel.model.ModelChannel;
import org.apache.camel.model.OnCompletionDefinition;
import org.apache.camel.model.OnExceptionDefinition;
import org.apache.camel.model.ProcessorDefinition;
import org.apache.camel.model.ProcessorDefinitionHelper;
import org.apache.camel.model.RouteDefinition;
import org.apache.camel.model.RouteDefinitionHelper;
import org.apache.camel.processor.CamelInternalProcessor;
import org.apache.camel.processor.InterceptorToAsyncProcessorBridge;
import org.apache.camel.processor.RedeliveryErrorHandler;
import org.apache.camel.processor.WrapProcessor;
import org.apache.camel.spi.InterceptStrategy;
import org.apache.camel.spi.MessageHistoryFactory;
import org.apache.camel.spi.RouteContext;
import org.apache.camel.util.OrderedComparator;
import org.apache.camel.util.ServiceHelper;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
/**
* DefaultChannel is the default {@link Channel}.
*
* The current implementation is just a composite containing the interceptors and error handler
* that beforehand was added to the route graph directly.
*
* With this {@link Channel} we can in the future implement better strategies for routing the
* {@link Exchange} in the route graph, as we have a {@link Channel} between each and every node
* in the graph.
*
* @version
*/
public class DefaultChannel extends CamelInternalProcessor implements ModelChannel {
private static final Logger LOG = LoggerFactory.getLogger(DefaultChannel.class);
private final List interceptors = new ArrayList<>();
private Processor errorHandler;
// the next processor (non wrapped)
private Processor nextProcessor;
// the real output to invoke that has been wrapped
private Processor output;
private ProcessorDefinition definition;
private ProcessorDefinition childDefinition;
private InstrumentationProcessor instrumentationProcessor;
private CamelContext camelContext;
private RouteContext routeContext;
public void setNextProcessor(Processor next) {
this.nextProcessor = next;
}
public Processor getOutput() {
// the errorHandler is already decorated with interceptors
// so it contain the entire chain of processors, so we can safely use it directly as output
// if no error handler provided we use the output
// TODO: Camel 3.0 we should determine the output dynamically at runtime instead of having the
// the error handlers, interceptors, etc. woven in at design time
return errorHandler != null ? errorHandler : output;
}
@Override
public boolean hasNext() {
return nextProcessor != null;
}
@Override
public List next() {
if (!hasNext()) {
return null;
}
List answer = new ArrayList<>(1);
answer.add(nextProcessor);
return answer;
}
public void setOutput(Processor output) {
this.output = output;
}
public Processor getNextProcessor() {
return nextProcessor;
}
public boolean hasInterceptorStrategy(Class type) {
for (InterceptStrategy strategy : interceptors) {
if (type.isInstance(strategy)) {
return true;
}
}
return false;
}
public void setErrorHandler(Processor errorHandler) {
this.errorHandler = errorHandler;
}
public Processor getErrorHandler() {
return errorHandler;
}
public void addInterceptStrategy(InterceptStrategy strategy) {
interceptors.add(strategy);
}
public void addInterceptStrategies(List strategies) {
interceptors.addAll(strategies);
}
public List getInterceptStrategies() {
return interceptors;
}
public ProcessorDefinition getProcessorDefinition() {
return definition;
}
public void setChildDefinition(ProcessorDefinition childDefinition) {
this.childDefinition = childDefinition;
}
public RouteContext getRouteContext() {
return routeContext;
}
@Override
protected void doStart() throws Exception {
// the output has now been created, so assign the output as the processor
setProcessor(getOutput());
ServiceHelper.startServices(errorHandler, output);
}
@Override
protected void doStop() throws Exception {
if (!isContextScoped()) {
// only stop services if not context scoped (as context scoped is reused by others)
ServiceHelper.stopServices(output, errorHandler);
}
}
@Override
protected void doShutdown() throws Exception {
ServiceHelper.stopAndShutdownServices(output, errorHandler);
}
private boolean isContextScoped() {
if (definition instanceof OnExceptionDefinition) {
return !((OnExceptionDefinition) definition).isRouteScoped();
} else if (definition instanceof OnCompletionDefinition) {
return !((OnCompletionDefinition) definition).isRouteScoped();
}
return false;
}
@SuppressWarnings("deprecation")
public void initChannel(ProcessorDefinition outputDefinition, RouteContext routeContext) throws Exception {
this.routeContext = routeContext;
this.definition = outputDefinition;
this.camelContext = routeContext.getCamelContext();
Processor target = nextProcessor;
Processor next;
// init CamelContextAware as early as possible on target
if (target instanceof CamelContextAware) {
((CamelContextAware) target).setCamelContext(camelContext);
}
// the definition to wrap should be the fine grained,
// so if a child is set then use it, if not then its the original output used
ProcessorDefinition targetOutputDef = childDefinition != null ? childDefinition : outputDefinition;
LOG.debug("Initialize channel for target: '{}'", targetOutputDef);
// fix parent/child relationship. This will be the case of the routes has been
// defined using XML DSL or end user may have manually assembled a route from the model.
// Background note: parent/child relationship is assembled on-the-fly when using Java DSL (fluent builders)
// where as when using XML DSL (JAXB) then it fixed after, but if people are using custom interceptors
// then we need to fix the parent/child relationship beforehand, and thus we can do it here
// ideally we need the design time route -> runtime route to be a 2-phase pass (scheduled work for Camel 3.0)
if (childDefinition != null && outputDefinition != childDefinition) {
childDefinition.setParent(outputDefinition);
}
// force the creation of an id
RouteDefinitionHelper.forceAssignIds(routeContext.getCamelContext(), definition);
// setup instrumentation processor for management (jmx)
// this is later used in postInitChannel as we need to setup the error handler later as well
InterceptStrategy managed = routeContext.getManagedInterceptStrategy();
if (managed instanceof InstrumentationInterceptStrategy) {
InstrumentationInterceptStrategy iis = (InstrumentationInterceptStrategy) managed;
instrumentationProcessor = new InstrumentationProcessor(targetOutputDef.getShortName(), target);
iis.prepareProcessor(targetOutputDef, target, instrumentationProcessor);
}
// then wrap the output with the backlog and tracer (backlog first, as we do not want regular tracer to trace the backlog)
InterceptStrategy tracer = getOrCreateBacklogTracer();
camelContext.addService(tracer);
if (tracer instanceof BacklogTracer) {
BacklogTracer backlogTracer = (BacklogTracer) tracer;
RouteDefinition route = ProcessorDefinitionHelper.getRoute(definition);
boolean first = false;
if (route != null && !route.getOutputs().isEmpty()) {
first = route.getOutputs().get(0) == definition;
}
addAdvice(new BacklogTracerAdvice(backlogTracer, targetOutputDef, route, first));
// add debugger as well so we have both tracing and debugging out of the box
InterceptStrategy debugger = getOrCreateBacklogDebugger();
camelContext.addService(debugger);
if (debugger instanceof BacklogDebugger) {
BacklogDebugger backlogDebugger = (BacklogDebugger) debugger;
addAdvice(new BacklogDebuggerAdvice(backlogDebugger, target, targetOutputDef));
}
}
if (routeContext.isMessageHistory()) {
// add message history advice
MessageHistoryFactory factory = camelContext.getMessageHistoryFactory();
addAdvice(new MessageHistoryAdvice(factory, targetOutputDef));
}
// the regular tracer is not a task on internalProcessor as this is not really needed
// end users have to explicit enable the tracer to use it, and then its okay if we wrap
// the processors (but by default tracer is disabled, and therefore we do not wrap processors)
tracer = getOrCreateTracer();
if (tracer != null) {
camelContext.addService(tracer);
TraceInterceptor trace = (TraceInterceptor) tracer.wrapProcessorInInterceptors(routeContext.getCamelContext(), targetOutputDef, target, null);
// trace interceptor need to have a reference to route context so we at runtime can enable/disable tracing on-the-fly
trace.setRouteContext(routeContext);
target = trace;
}
// sort interceptors according to ordered
interceptors.sort(OrderedComparator.get());
// then reverse list so the first will be wrapped last, as it would then be first being invoked
Collections.reverse(interceptors);
// wrap the output with the configured interceptors
for (InterceptStrategy strategy : interceptors) {
next = target == nextProcessor ? null : nextProcessor;
// skip tracer as we did the specially beforehand and it could potentially be added as an interceptor strategy
if (strategy instanceof Tracer) {
continue;
}
// skip stream caching as it must be wrapped as outer most, which we do later
if (strategy instanceof StreamCaching) {
continue;
}
// use the fine grained definition (eg the child if available). Its always possible to get back to the parent
Processor wrapped = strategy.wrapProcessorInInterceptors(routeContext.getCamelContext(), targetOutputDef, target, next);
if (!(wrapped instanceof AsyncProcessor)) {
LOG.warn("Interceptor: " + strategy + " at: " + outputDefinition + " does not return an AsyncProcessor instance."
+ " This causes the asynchronous routing engine to not work as optimal as possible."
+ " See more details at the InterceptStrategy javadoc."
+ " Camel will use a bridge to adapt the interceptor to the asynchronous routing engine,"
+ " but its not the most optimal solution. Please consider changing your interceptor to comply.");
// use a bridge and wrap again which allows us to adapt and leverage the asynchronous routing engine anyway
// however its not the most optimal solution, but we can still run.
InterceptorToAsyncProcessorBridge bridge = new InterceptorToAsyncProcessorBridge(target);
wrapped = strategy.wrapProcessorInInterceptors(routeContext.getCamelContext(), targetOutputDef, bridge, next);
// Avoid the stack overflow
if (!wrapped.equals(bridge)) {
bridge.setTarget(wrapped);
} else {
// Just skip the wrapped processor
bridge.setTarget(null);
}
wrapped = bridge;
}
if (!(wrapped instanceof WrapProcessor)) {
// wrap the target so it becomes a service and we can manage its lifecycle
wrapped = new WrapProcessor(wrapped, target);
}
target = wrapped;
}
if (routeContext.isStreamCaching()) {
addAdvice(new StreamCachingAdvice(camelContext.getStreamCachingStrategy()));
}
if (routeContext.getDelayer() != null && routeContext.getDelayer() > 0) {
addAdvice(new DelayerAdvice(routeContext.getDelayer()));
}
// sets the delegate to our wrapped output
output = target;
}
@Override
public void postInitChannel(ProcessorDefinition outputDefinition, RouteContext routeContext) throws Exception {
// if jmx was enabled for the processor then either add as advice or wrap and change the processor
// on the error handler. See more details in the class javadoc of InstrumentationProcessor
if (instrumentationProcessor != null) {
boolean redeliveryPossible = false;
if (errorHandler instanceof RedeliveryErrorHandler) {
redeliveryPossible = ((RedeliveryErrorHandler) errorHandler).determineIfRedeliveryIsEnabled();
if (redeliveryPossible) {
// okay we can redeliver then we need to change the output in the error handler
// to use us which we then wrap the call so we can capture before/after for redeliveries as well
Processor currentOutput = ((RedeliveryErrorHandler) errorHandler).getOutput();
instrumentationProcessor.setProcessor(currentOutput);
((RedeliveryErrorHandler) errorHandler).changeOutput(instrumentationProcessor);
}
}
if (!redeliveryPossible) {
// optimise to use advice as we cannot redeliver
addAdvice(instrumentationProcessor);
}
}
}
private InterceptStrategy getOrCreateTracer() {
// only use tracer if explicit enabled
if (camelContext.isTracing() != null && !camelContext.isTracing()) {
return null;
}
InterceptStrategy tracer = Tracer.getTracer(camelContext);
if (tracer == null) {
if (camelContext.getRegistry() != null) {
// lookup in registry
Map map = camelContext.getRegistry().findByTypeWithName(Tracer.class);
if (map.size() == 1) {
tracer = map.values().iterator().next();
}
}
if (tracer == null) {
// fallback to use the default tracer
tracer = camelContext.getDefaultTracer();
// configure and use any trace formatter if any exists
Map formatters = camelContext.getRegistry().findByTypeWithName(TraceFormatter.class);
if (formatters.size() == 1) {
TraceFormatter formatter = formatters.values().iterator().next();
if (tracer instanceof Tracer) {
((Tracer) tracer).setFormatter(formatter);
}
}
}
}
return tracer;
}
private InterceptStrategy getOrCreateBacklogTracer() {
InterceptStrategy tracer = BacklogTracer.getBacklogTracer(camelContext);
if (tracer == null) {
if (camelContext.getRegistry() != null) {
// lookup in registry
Map map = camelContext.getRegistry().findByTypeWithName(BacklogTracer.class);
if (map.size() == 1) {
tracer = map.values().iterator().next();
}
}
if (tracer == null) {
// fallback to use the default tracer
tracer = camelContext.getDefaultBacklogTracer();
}
}
return tracer;
}
private InterceptStrategy getOrCreateBacklogDebugger() {
InterceptStrategy debugger = BacklogDebugger.getBacklogDebugger(camelContext);
if (debugger == null) {
if (camelContext.getRegistry() != null) {
// lookup in registry
Map map = camelContext.getRegistry().findByTypeWithName(BacklogDebugger.class);
if (map.size() == 1) {
debugger = map.values().iterator().next();
}
}
if (debugger == null) {
// fallback to use the default debugger
debugger = camelContext.getDefaultBacklogDebugger();
}
}
return debugger;
}
@Override
public String toString() {
// just output the next processor as all the interceptors and error handler is just too verbose
return "Channel[" + nextProcessor + "]";
}
}