SkyWalking链路追踪上下文TraceContext的追踪身份traceId生成的实现原理剖析

结论先行

SkyWalking 通过字节码增强技术实现,结合依赖注入和控制反转思想,以SkyWalking方式将追踪身份traceId编织到链路追踪上下文TraceContext中。
是不是很有趣,很有意思!!!

实现原理剖析

org.apache.skywalking:java-agent:9.1.0
v9.1.0

TraceContext.traceId()

org.apache.skywalking.apm.toolkit.trace.TraceContext#traceId

请求链路追踪上下文TraceContext,调用TraceContext.traceId()获取追踪身份traceId

package org.apache.skywalking.apm.toolkit.trace;import java.util.Optional;/*** Try to access the sky-walking tracer context. The context is not existed, always. only the middleware, component, or* rpc-framework are supported in the current invoke stack, in the same thread, the context will be available.* <p>*/
public class TraceContext {/*** Try to get the traceId of current trace context.* 尝试获取当前追踪上下文的追踪身份traceId** @return traceId, if it exists, or empty {@link String}.*/public static String traceId() {return "";}/*** Try to get the segmentId of current trace context.** @return segmentId, if it exists, or empty {@link String}.*/public static String segmentId() {return "";}/*** Try to get the spanId of current trace context. The spanId is a negative number when the trace context is* missing.** @return spanId, if it exists, or empty {@link String}.*/public static int spanId() {return -1;}/*** Try to get the custom value from trace context.** @return custom data value.*/public static Optional<String> getCorrelation(String key) {return Optional.empty();}/*** Put the custom key/value into trace context.** @return previous value if it exists.*/public static Optional<String> putCorrelation(String key, String value) {return Optional.empty();}}

1.链路追踪上下文的traceId是如何设置进去的?

skywalking:java-agent项目仓库里搜索org.apache.skywalking.apm.toolkit.trace.TraceContext
repo:apache/skywalking-java org.apache.skywalking.apm.toolkit.trace.TraceContext language:Java
在这里插入图片描述
在IDEA skywalking:java-agent项目源代码里搜索org.apache.skywalking.apm.toolkit.trace.TraceContext
在这里插入图片描述

【结论】通过字节码增强技术实现,结合依赖注入和控制反转思想,以SkyWalking方式将追踪身份traceId编织到链路追踪上下文TraceContext中。
在这里插入图片描述

TraceContextActivation

org.apache.skywalking.apm.toolkit.activation.trace.TraceContextActivation

链路追踪上下文激活TraceContextActivation,通过TraceIDInterceptor拦截TraceContext.traceId(),将追踪身份traceId设置到链路追踪上下文TraceContext

package org.apache.skywalking.apm.toolkit.activation.trace;import net.bytebuddy.description.method.MethodDescription;
import net.bytebuddy.matcher.ElementMatcher;
import org.apache.skywalking.apm.agent.core.plugin.interceptor.enhance.ClassStaticMethodsEnhancePluginDefine;
import org.apache.skywalking.apm.agent.core.plugin.match.ClassMatch;
import org.apache.skywalking.apm.agent.core.plugin.match.NameMatch;
import org.apache.skywalking.apm.agent.core.plugin.interceptor.StaticMethodsInterceptPoint;import static net.bytebuddy.matcher.ElementMatchers.named;/*** Active the toolkit class "TraceContext". Should not dependency or import any class in* "skywalking-toolkit-trace-context" module. Activation's classloader is diff from "TraceContext", using direct will* trigger classloader issue.* <p>*/
public class TraceContextActivation extends ClassStaticMethodsEnhancePluginDefine {// 追踪身份traceId拦截类public static final String TRACE_ID_INTERCEPT_CLASS = "org.apache.skywalking.apm.toolkit.activation.trace.TraceIDInterceptor";public static final String SEGMENT_ID_INTERCEPT_CLASS = "org.apache.skywalking.apm.toolkit.activation.trace.SegmentIDInterceptor";public static final String SPAN_ID_INTERCEPT_CLASS = "org.apache.skywalking.apm.toolkit.activation.trace.SpanIDInterceptor";// 增强类-追踪上下文public static final String ENHANCE_CLASS = "org.apache.skywalking.apm.toolkit.trace.TraceContext";// 获取追踪身份traceId的静态方法名称public static final String ENHANCE_TRACE_ID_METHOD = "traceId";public static final String ENHANCE_SEGMENT_ID_METHOD = "segmentId";public static final String ENHANCE_SPAN_ID_METHOD = "spanId";public static final String ENHANCE_GET_CORRELATION_METHOD = "getCorrelation";public static final String INTERCEPT_GET_CORRELATION_CLASS = "org.apache.skywalking.apm.toolkit.activation.trace.CorrelationContextGetInterceptor";public static final String ENHANCE_PUT_CORRELATION_METHOD = "putCorrelation";public static final String INTERCEPT_PUT_CORRELATION_CLASS = "org.apache.skywalking.apm.toolkit.activation.trace.CorrelationContextPutInterceptor";/*** @return the target class, which needs active.*/@Overrideprotected ClassMatch enhanceClass() {// 增强类return NameMatch.byName(ENHANCE_CLASS);}/*** @return the collection of {@link StaticMethodsInterceptPoint}, represent the intercepted methods and their* interceptors.*/@Overridepublic StaticMethodsInterceptPoint[] getStaticMethodsInterceptPoints() {// 静态方法拦截点return new StaticMethodsInterceptPoint[] {new StaticMethodsInterceptPoint() {@Overridepublic ElementMatcher<MethodDescription> getMethodsMatcher() {// 获取追踪身份traceId的静态方法名称return named(ENHANCE_TRACE_ID_METHOD);}@Overridepublic String getMethodsInterceptor() {// 追踪身份traceId拦截类return TRACE_ID_INTERCEPT_CLASS;}@Overridepublic boolean isOverrideArgs() {return false;}},// ...};}
}

TraceIDInterceptor

org.apache.skywalking.apm.toolkit.activation.trace.TraceIDInterceptor

追踪身份拦截器TraceIDInterceptor,调用ContextManager.getGlobalTraceId()获取追踪身份traceId,将其返回给TraceContext.traceId()
在这里插入图片描述

package org.apache.skywalking.apm.toolkit.activation.trace;import java.lang.reflect.Method;
import org.apache.skywalking.apm.agent.core.logging.api.ILog;
import org.apache.skywalking.apm.agent.core.logging.api.LogManager;
import org.apache.skywalking.apm.agent.core.plugin.interceptor.enhance.StaticMethodsAroundInterceptor;
import org.apache.skywalking.apm.agent.core.context.ContextManager;
import org.apache.skywalking.apm.agent.core.plugin.interceptor.enhance.MethodInterceptResult;public class TraceIDInterceptor implements StaticMethodsAroundInterceptor {private static final ILog LOGGER = LogManager.getLogger(TraceIDInterceptor.class);@Overridepublic void beforeMethod(Class clazz, Method method, Object[] allArguments, Class<?>[] parameterTypes,MethodInterceptResult result) {// 获取第一个全局追踪身份traceId,将其定义为方法返回值result.defineReturnValue(ContextManager.getGlobalTraceId());}@Overridepublic Object afterMethod(Class clazz, Method method, Object[] allArguments, Class<?>[] parameterTypes,Object ret) {// 返回追踪身份traceIdreturn ret;}@Overridepublic void handleMethodException(Class clazz, Method method, Object[] allArguments, Class<?>[] parameterTypes,Throwable t) {LOGGER.error("Failed to getDefault trace Id.", t);}
}

ContextManager.getGlobalTraceId()

org.apache.skywalking.apm.agent.core.context.ContextManager#getGlobalTraceId

链路追踪上下文管理器ContextManager
ContextManager.getGlobalTraceId()是获取第一个全局追踪身份traceId,其调用AbstractTracerContext.getReadablePrimaryTraceId()获取全局追踪身份traceId
在这里插入图片描述

package org.apache.skywalking.apm.agent.core.context;import java.util.Objects;
import org.apache.skywalking.apm.agent.core.boot.BootService;
import org.apache.skywalking.apm.agent.core.boot.ServiceManager;
import org.apache.skywalking.apm.agent.core.context.trace.AbstractSpan;
import org.apache.skywalking.apm.agent.core.context.trace.TraceSegment;
import org.apache.skywalking.apm.agent.core.logging.api.ILog;
import org.apache.skywalking.apm.agent.core.logging.api.LogManager;
import org.apache.skywalking.apm.agent.core.sampling.SamplingService;
import org.apache.skywalking.apm.util.StringUtil;import static org.apache.skywalking.apm.agent.core.conf.Config.Agent.OPERATION_NAME_THRESHOLD;/*** {@link ContextManager} controls the whole context of {@link TraceSegment}. Any {@link TraceSegment} relates to* single-thread, so this context use {@link ThreadLocal} to maintain the context, and make sure, since a {@link* TraceSegment} starts, all ChildOf spans are in the same context. <p> What is 'ChildOf'?* https://github.com/opentracing/specification/blob/master/specification.md#references-between-spans** <p> Also, {@link ContextManager} delegates to all {@link AbstractTracerContext}'s major methods.*/
public class ContextManager implements BootService {private static final String EMPTY_TRACE_CONTEXT_ID = "N/A";private static final ILog LOGGER = LogManager.getLogger(ContextManager.class);// 追踪上下文的线程本地变量private static ThreadLocal<AbstractTracerContext> CONTEXT = new ThreadLocal<AbstractTracerContext>();private static ThreadLocal<RuntimeContext> RUNTIME_CONTEXT = new ThreadLocal<RuntimeContext>();private static ContextManagerExtendService EXTEND_SERVICE;private static AbstractTracerContext getOrCreate(String operationName, boolean forceSampling) {AbstractTracerContext context = CONTEXT.get();if (context == null) {if (StringUtil.isEmpty(operationName)) {if (LOGGER.isDebugEnable()) {LOGGER.debug("No operation name, ignore this trace.");}context = new IgnoredTracerContext();} else {if (EXTEND_SERVICE == null) {EXTEND_SERVICE = ServiceManager.INSTANCE.findService(ContextManagerExtendService.class);}context = EXTEND_SERVICE.createTraceContext(operationName, forceSampling);}CONTEXT.set(context);}return context;}/*** 获取第一个全局追踪身份traceId* @return the first global trace id when tracing. Otherwise, "N/A".*/public static String getGlobalTraceId() {// 追踪上下文AbstractTracerContext context = CONTEXT.get();// 获取全局追踪身份traceIdreturn Objects.nonNull(context) ? context.getReadablePrimaryTraceId() : EMPTY_TRACE_CONTEXT_ID;}/*** @return the current segment id when tracing. Otherwise, "N/A".*/public static String getSegmentId() {AbstractTracerContext context = CONTEXT.get();return Objects.nonNull(context) ? context.getSegmentId() : EMPTY_TRACE_CONTEXT_ID;}/*** @return the current span id when tracing. Otherwise, the value is -1.*/public static int getSpanId() {AbstractTracerContext context = CONTEXT.get();return Objects.nonNull(context) ? context.getSpanId() : -1;}// ...}

AbstractTracerContext.getReadablePrimaryTraceId()

org.apache.skywalking.apm.agent.core.context.AbstractTracerContext#getReadablePrimaryTraceId

追踪上下文定义接口AbstractTracerContext
本方法获取全局追踪身份traceId

package org.apache.skywalking.apm.agent.core.context;import org.apache.skywalking.apm.agent.core.context.trace.AbstractSpan;/*** The <code>AbstractTracerContext</code> represents the tracer context manager.* 表示追踪上下文管理器*/
public interface AbstractTracerContext {/*** Get the global trace id, if needEnhance. How to build, depends on the implementation.* 获取全局追踪身份traceId** @return the string represents the id.*/String getReadablePrimaryTraceId();/*** Prepare for the cross-process propagation. How to initialize the carrier, depends on the implementation.** @param carrier to carry the context for crossing process.*/void inject(ContextCarrier carrier);/*** Build the reference between this segment and a cross-process segment. How to build, depends on the* implementation.** @param carrier carried the context from a cross-process segment.*/void extract(ContextCarrier carrier);/*** Capture a snapshot for cross-thread propagation. It's a similar concept with ActiveSpan.Continuation in* OpenTracing-java How to build, depends on the implementation.** @return the {@link ContextSnapshot} , which includes the reference context.*/ContextSnapshot capture();/*** Build the reference between this segment and a cross-thread segment. How to build, depends on the* implementation.** @param snapshot from {@link #capture()} in the parent thread.*/void continued(ContextSnapshot snapshot);/*** Get the current segment id, if needEnhance. How to build, depends on the implementation.** @return the string represents the id.*/String getSegmentId();/*** Get the active span id, if needEnhance. How to build, depends on the implementation.** @return the string represents the id.*/int getSpanId();/*** Create an entry span** @param operationName most likely a service name* @return the span represents an entry point of this segment.*/AbstractSpan createEntrySpan(String operationName);/*** Create a local span** @param operationName most likely a local method signature, or business name.* @return the span represents a local logic block.*/AbstractSpan createLocalSpan(String operationName);/*** Create an exit span** @param operationName most likely a service name of remote* @param remotePeer    the network id(ip:port, hostname:port or ip1:port1,ip2,port, etc.). Remote peer could be set*                      later, but must be before injecting.* @return the span represent an exit point of this segment.*/AbstractSpan createExitSpan(String operationName, String remotePeer);/*** @return the active span of current tracing context(stack)*/AbstractSpan activeSpan();/*** Finish the given span, and the given span should be the active span of current tracing context(stack)** @param span to finish* @return true when context should be clear.*/boolean stopSpan(AbstractSpan span);/*** Notify this context, current span is going to be finished async in another thread.** @return The current context*/AbstractTracerContext awaitFinishAsync();/*** The given span could be stopped officially.** @param span to be stopped.*/void asyncStop(AsyncSpan span);/*** Get current correlation context*/CorrelationContext getCorrelationContext();/*** Get current primary endpoint name*/String getPrimaryEndpointName();
}

AbstractTracerContext有两个子类IgnoredTracerContextTracingContext

IgnoredTracerContext.getReadablePrimaryTraceId()

org.apache.skywalking.apm.agent.core.context.IgnoredTracerContext#getReadablePrimaryTraceId

可忽略的追踪上下文IgnoredTracerContext
本方法返回"Ignored_Trace"

package org.apache.skywalking.apm.agent.core.context;import java.util.LinkedList;
import java.util.List;
import org.apache.skywalking.apm.agent.core.context.trace.AbstractSpan;
import org.apache.skywalking.apm.agent.core.context.trace.NoopSpan;
import org.apache.skywalking.apm.agent.core.profile.ProfileStatusContext;/*** The <code>IgnoredTracerContext</code> represent a context should be ignored. So it just maintains the stack with an* integer depth field.* <p>* All operations through this will be ignored, and keep the memory and gc cost as low as possible.*/
public class IgnoredTracerContext implements AbstractTracerContext {private static final NoopSpan NOOP_SPAN = new NoopSpan();private static final String IGNORE_TRACE = "Ignored_Trace";private final CorrelationContext correlationContext;private final ExtensionContext extensionContext;private final ProfileStatusContext profileStatusContext;private int stackDepth;public IgnoredTracerContext() {this.stackDepth = 0;this.correlationContext = new CorrelationContext();this.extensionContext = new ExtensionContext();this.profileStatusContext = ProfileStatusContext.createWithNone();}// ...@Overridepublic String getReadablePrimaryTraceId() {// 获取全局追踪身份traceIdreturn IGNORE_TRACE;}@Overridepublic String getSegmentId() {return IGNORE_TRACE;}@Overridepublic int getSpanId() {return -1;}// ...}

TracingContext.getReadablePrimaryTraceId()

org.apache.skywalking.apm.agent.core.context.TracingContext#getReadablePrimaryTraceId

链路追踪上下文TracingContext
本方法返回DistributedTraceIdid字段属性
在这里插入图片描述

package org.apache.skywalking.apm.agent.core.context;import java.util.LinkedList;
import java.util.List;
import java.util.concurrent.atomic.AtomicIntegerFieldUpdater;
import java.util.concurrent.locks.ReentrantLock;
import org.apache.skywalking.apm.agent.core.boot.ServiceManager;
import org.apache.skywalking.apm.agent.core.conf.Config;
import org.apache.skywalking.apm.agent.core.conf.dynamic.watcher.SpanLimitWatcher;
import org.apache.skywalking.apm.agent.core.context.ids.DistributedTraceId;
import org.apache.skywalking.apm.agent.core.context.ids.PropagatedTraceId;
import org.apache.skywalking.apm.agent.core.context.trace.AbstractSpan;
import org.apache.skywalking.apm.agent.core.context.trace.AbstractTracingSpan;
import org.apache.skywalking.apm.agent.core.context.trace.EntrySpan;
import org.apache.skywalking.apm.agent.core.context.trace.ExitSpan;
import org.apache.skywalking.apm.agent.core.context.trace.ExitTypeSpan;
import org.apache.skywalking.apm.agent.core.context.trace.LocalSpan;
import org.apache.skywalking.apm.agent.core.context.trace.NoopExitSpan;
import org.apache.skywalking.apm.agent.core.context.trace.NoopSpan;
import org.apache.skywalking.apm.agent.core.context.trace.TraceSegment;
import org.apache.skywalking.apm.agent.core.context.trace.TraceSegmentRef;
import org.apache.skywalking.apm.agent.core.logging.api.ILog;
import org.apache.skywalking.apm.agent.core.logging.api.LogManager;
import org.apache.skywalking.apm.agent.core.profile.ProfileStatusContext;
import org.apache.skywalking.apm.agent.core.profile.ProfileTaskExecutionService;
import org.apache.skywalking.apm.util.StringUtil;import static org.apache.skywalking.apm.agent.core.conf.Config.Agent.CLUSTER;/*** The <code>TracingContext</code> represents a core tracing logic controller. It build the final {@link* TracingContext}, by the stack mechanism, which is similar with the codes work.* <p>* In opentracing concept, it means, all spans in a segment tracing context(thread) are CHILD_OF relationship, but no* FOLLOW_OF.* <p>* In skywalking core concept, FOLLOW_OF is an abstract concept when cross-process MQ or cross-thread async/batch tasks* happen, we used {@link TraceSegmentRef} for these scenarios. Check {@link TraceSegmentRef} which is from {@link* ContextCarrier} or {@link ContextSnapshot}.*/
public class TracingContext implements AbstractTracerContext {private static final ILog LOGGER = LogManager.getLogger(TracingContext.class);private long lastWarningTimestamp = 0;/*** @see ProfileTaskExecutionService*/private static ProfileTaskExecutionService PROFILE_TASK_EXECUTION_SERVICE;/*** The final {@link TraceSegment}, which includes all finished spans.* 追踪段,同一线程内的所有调用*/private TraceSegment segment;/*** Active spans stored in a Stack, usually called 'ActiveSpanStack'. This {@link LinkedList} is the in-memory* storage-structure. <p> I use {@link LinkedList#removeLast()}, {@link LinkedList#addLast(Object)} and {@link* LinkedList#getLast()} instead of {@link #pop()}, {@link #push(AbstractSpan)}, {@link #peek()}*/private LinkedList<AbstractSpan> activeSpanStack = new LinkedList<>();/*** @since 8.10.0 replace the removed "firstSpan"(before 8.10.0) reference. see {@link PrimaryEndpoint} for more details.*/private PrimaryEndpoint primaryEndpoint = null;/*** A counter for the next span.*/private int spanIdGenerator;/*** The counter indicates*/@SuppressWarnings("unused") // updated by ASYNC_SPAN_COUNTER_UPDATERprivate volatile int asyncSpanCounter;private static final AtomicIntegerFieldUpdater<TracingContext> ASYNC_SPAN_COUNTER_UPDATER =AtomicIntegerFieldUpdater.newUpdater(TracingContext.class, "asyncSpanCounter");private volatile boolean isRunningInAsyncMode;private volatile ReentrantLock asyncFinishLock;private volatile boolean running;private final long createTime;/*** profile status*/private final ProfileStatusContext profileStatus;@Getter(AccessLevel.PACKAGE)private final CorrelationContext correlationContext;@Getter(AccessLevel.PACKAGE)private final ExtensionContext extensionContext;//CDS watcherprivate final SpanLimitWatcher spanLimitWatcher;/*** Initialize all fields with default value.*/TracingContext(String firstOPName, SpanLimitWatcher spanLimitWatcher) {this.segment = new TraceSegment();this.spanIdGenerator = 0;isRunningInAsyncMode = false;createTime = System.currentTimeMillis();running = true;// profiling statusif (PROFILE_TASK_EXECUTION_SERVICE == null) {PROFILE_TASK_EXECUTION_SERVICE = ServiceManager.INSTANCE.findService(ProfileTaskExecutionService.class);}this.profileStatus = PROFILE_TASK_EXECUTION_SERVICE.addProfiling(this, segment.getTraceSegmentId(), firstOPName);this.correlationContext = new CorrelationContext();this.extensionContext = new ExtensionContext();this.spanLimitWatcher = spanLimitWatcher;}/*** 获取全局追踪身份traceId* @return the first global trace id.*/@Overridepublic String getReadablePrimaryTraceId() {// 获取分布式的追踪身份的id字段属性return getPrimaryTraceId().getId();}private DistributedTraceId getPrimaryTraceId() {// 获取追踪段相关的分布式的追踪身份return segment.getRelatedGlobalTrace();}@Overridepublic String getSegmentId() {return segment.getTraceSegmentId();}@Overridepublic int getSpanId() {return activeSpan().getSpanId();}// ...}

DistributedTraceId

org.apache.skywalking.apm.agent.core.context.ids.DistributedTraceId#id

分布式的追踪身份DistributedTraceId,表示一个分布式调用链路。

package org.apache.skywalking.apm.agent.core.context.ids;import lombok.EqualsAndHashCode;
import lombok.Getter;
import lombok.RequiredArgsConstructor;
import lombok.ToString;/*** The <code>DistributedTraceId</code> presents a distributed call chain.* 表示一个分布式调用链路。* <p>* This call chain has a unique (service) entrance,* <p>* such as: Service : http://www.skywalking.com/cust/query, all the remote, called behind this service, rest remote, db* executions, are using the same <code>DistributedTraceId</code> even in different JVM.* <p>* The <code>DistributedTraceId</code> contains only one string, and can NOT be reset, creating a new instance is the* only option.*/
@RequiredArgsConstructor
@ToString
@EqualsAndHashCode
public abstract class DistributedTraceId {@Getterprivate final String id;
}

DistributedTraceId有两个子类PropagatedTraceIdNewDistributedTraceId

PropagatedTraceId

org.apache.skywalking.apm.agent.core.context.ids.PropagatedTraceId

传播的追踪身份PropagatedTraceId,表示从对等端传播的DistributedTraceId
在这里插入图片描述

package org.apache.skywalking.apm.agent.core.context.ids;/*** The <code>PropagatedTraceId</code> represents a {@link DistributedTraceId}, which is propagated from the peer.*/
public class PropagatedTraceId extends DistributedTraceId {public PropagatedTraceId(String id) {// 透传追踪身份traceIdsuper(id);}
}

NewDistributedTraceId

org.apache.skywalking.apm.agent.core.context.ids.NewDistributedTraceId

新的分布式的追踪身份NewDistributedTraceId,是具有新生成的idDistributedTraceId
默认构造函数调用GlobalIdGenerator.generate()生成新的全局id,即追踪身份traceId
在这里插入图片描述

package org.apache.skywalking.apm.agent.core.context.ids;/*** The <code>NewDistributedTraceId</code> is a {@link DistributedTraceId} with a new generated id.*/
public class NewDistributedTraceId extends DistributedTraceId {public NewDistributedTraceId() {生成新的全局id,即追踪身份traceIdsuper(GlobalIdGenerator.generate());}
}

GlobalIdGenerator.generate()

org.apache.skywalking.apm.agent.core.context.ids.GlobalIdGenerator#generate

全局id生成器GlobalIdGenerator
本方法用于生成一个新的全局id,是真正生成追踪身份traceId的地方。
在这里插入图片描述

package org.apache.skywalking.apm.agent.core.context.ids;import java.util.UUID;import org.apache.skywalking.apm.util.StringUtil;public final class GlobalIdGenerator {// 应用实例进程身份idprivate static final String PROCESS_ID = UUID.randomUUID().toString().replaceAll("-", "");// 线程的id序列号的上下文private static final ThreadLocal<IDContext> THREAD_ID_SEQUENCE = ThreadLocal.withInitial(() -> new IDContext(System.currentTimeMillis(), (short) 0));private GlobalIdGenerator() {}/*** 生成一个新的id。* Generate a new id, combined by three parts.* <p>* The first one represents application instance id.* 第一部分,表示应用实例进程身份id* <p>* The second one represents thread id.* 第二部分,表示线程身份id* <p>* The third one also has two parts, 1) a timestamp, measured in milliseconds 2) a seq, in current thread, between* 0(included) and 9999(included)* 第三部分,也有两个部分, 1) 一个时间戳,单位是毫秒ms 2) 在当前线程中的一个序列号,位于[0,9999]之间** @return unique id to represent a trace or segment* 表示追踪或追踪段的唯一id*/public static String generate() {return StringUtil.join('.',PROCESS_ID,String.valueOf(Thread.currentThread().getId()),String.valueOf(THREAD_ID_SEQUENCE.get().nextSeq()));}private static class IDContext {private long lastTimestamp;private short threadSeq;// Just for considering time-shift-back only.private long lastShiftTimestamp;private int lastShiftValue;private IDContext(long lastTimestamp, short threadSeq) {this.lastTimestamp = lastTimestamp;this.threadSeq = threadSeq;}private long nextSeq() {return timestamp() * 10000 + nextThreadSeq();}private long timestamp() {long currentTimeMillis = System.currentTimeMillis();if (currentTimeMillis < lastTimestamp) {// Just for considering time-shift-back by Ops or OS. @hanahmily 's suggestion.if (lastShiftTimestamp != currentTimeMillis) {lastShiftValue++;lastShiftTimestamp = currentTimeMillis;}return lastShiftValue;} else {lastTimestamp = currentTimeMillis;return lastTimestamp;}}private short nextThreadSeq() {if (threadSeq == 10000) {threadSeq = 0;}return threadSeq++;}}
}

示例验证

Arthas monitor/watch/trace 相关
若不了解其实现原理,是很难想到这个切面点。
实践出真知识!!!

stack org.apache.skywalking.apm.agent.core.context.ids.NewDistributedTraceId <init> '@java.lang.Thread@currentThread().getName().startsWith("wanda_event")'watch org.apache.skywalking.apm.agent.core.context.ids.NewDistributedTraceId <init> '{target, returnObj}' '@java.lang.Thread@currentThread().getName().startsWith("wanda_event")' -x 6
[arthas@7]$ stack org.apache.skywalking.apm.agent.core.context.AbstractTracerContext getReadablePrimaryTraceId '@java.lang.Thread@currentThread().getName().startsWith("wanda_event")'
Press Q or Ctrl+C to abort.
Affect(class count: 2 , method count: 1) cost in 423 ms, listenerId: 3
ts=2024-03-04 21:03:59;thread_name=wanda_event-thread-1;id=140;is_daemon=false;priority=5;TCCL=org.springframework.boot.loader.LaunchedURLClassLoader@67fe380b@org.apache.skywalking.apm.agent.core.context.TracingContext.getReadablePrimaryTraceId()at org.apache.skywalking.apm.agent.core.context.ContextManager.getGlobalTraceId(ContextManager.java:77)at org.apache.skywalking.apm.toolkit.activation.trace.TraceIDInterceptor.beforeMethod(TraceIDInterceptor.java:35)at org.apache.skywalking.apm.agent.core.plugin.interceptor.enhance.StaticMethodsInter.intercept(StaticMethodsInter.java:73)at org.apache.skywalking.apm.toolkit.trace.TraceContext.traceId(TraceContext.java:-1)// SkyWalking核心链路是上面👆🏻// 调用TraceContext.traceId()at com.leoao.lpaas.logback.LeoaoJsonLayout.addCustomDataToJsonMap(LeoaoJsonLayout.java:29)at ch.qos.logback.contrib.json.classic.JsonLayout.toJsonMap(null:-1)at ch.qos.logback.contrib.json.classic.JsonLayout.toJsonMap(null:-1)at ch.qos.logback.contrib.json.JsonLayoutBase.doLayout(null:-1)at ch.qos.logback.core.encoder.LayoutWrappingEncoder.encode(LayoutWrappingEncoder.java:115)at ch.qos.logback.core.OutputStreamAppender.subAppend(OutputStreamAppender.java:230)at ch.qos.logback.core.rolling.RollingFileAppender.subAppend(RollingFileAppender.java:235)at ch.qos.logback.core.OutputStreamAppender.append(OutputStreamAppender.java:102)at ch.qos.logback.core.UnsynchronizedAppenderBase.doAppend(UnsynchronizedAppenderBase.java:84)at ch.qos.logback.core.spi.AppenderAttachableImpl.appendLoopOnAppenders(AppenderAttachableImpl.java:51)at ch.qos.logback.classic.Logger.appendLoopOnAppenders(Logger.java:270)at ch.qos.logback.classic.Logger.callAppenders(Logger.java:257)at ch.qos.logback.classic.Logger.buildLoggingEventAndAppend(Logger.java:421)at ch.qos.logback.classic.Logger.filterAndLog_1(Logger.java:398)at ch.qos.logback.classic.Logger.info(Logger.java:583)// 输出打印日志// log.info("receive event persistUserPositionEvent=[{}]", event);at com.lefit.wanda.domain.event.listener.PersistUserPositionEventListener.change(PersistUserPositionEventListener.java:23)at sun.reflect.NativeMethodAccessorImpl.invoke0(NativeMethodAccessorImpl.java:-2)at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)at java.lang.reflect.Method.invoke(Method.java:498)at com.google.common.eventbus.Subscriber.invokeSubscriberMethod$original$ToNcZpNk(Subscriber.java:88)at com.google.common.eventbus.Subscriber.invokeSubscriberMethod$original$ToNcZpNk$accessor$utMvob4N(Subscriber.java:-1)at com.google.common.eventbus.Subscriber$auxiliary$8fYqzzq0.call(null:-1)at org.apache.skywalking.apm.agent.core.plugin.interceptor.enhance.InstMethodsInterWithOverrideArgs.intercept(InstMethodsInterWithOverrideArgs.java:85)at com.google.common.eventbus.Subscriber.invokeSubscriberMethod(Subscriber.java:-1)at com.google.common.eventbus.Subscriber$SynchronizedSubscriber.invokeSubscriberMethod(Subscriber.java:145)at com.google.common.eventbus.Subscriber$1.run(Subscriber.java:73)at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)at java.lang.Thread.run(Thread.java:750)

参考引用


祝大家玩得开心!ˇˍˇ

简放,杭州

本文来自互联网用户投稿,该文观点仅代表作者本人,不代表本站立场。本站仅提供信息存储空间服务,不拥有所有权,不承担相关法律责任。如若转载,请注明出处:http://www.rhkb.cn/news/269177.html

如若内容造成侵权/违法违规/事实不符,请联系长河编程网进行投诉反馈email:809451989@qq.com,一经查实,立即删除!

相关文章

B084-SpringCloud-Zuul Config

目录 zuul系统架构和zuul的作用zuul网关实现配置映射路径过滤器 Config概述云端管理本地配置 zuul zuul是分布式和集群后前端统一访问入口 系统架构和zuul的作用 zuul把自己注册进eureka&#xff0c;然后可通过前端传来的服务名发现和访问对应的服务集群 为了预防zuul单点故…

性能问题分析排查思路之机器(3)

本文是性能问题分析排查思路的展开内容之一&#xff0c;第2篇&#xff0c;主要分为日志1期&#xff0c;机器4期、环境2期共7篇系列文章&#xff0c;本期是第三篇&#xff0c;讲机器&#xff08;硬件&#xff09;的网络方面的排查方法和最佳实践。 主要内容如图所示&#xff1a…

Re61:读论文 PRP Get an A in Math: Progressive Rectification Prompting

诸神缄默不语-个人CSDN博文目录 诸神缄默不语的论文阅读笔记和分类 论文名称&#xff1a;Get an A in Math: Progressive Rectification Prompting ArXiv网址&#xff1a;https://arxiv.org/abs/2312.06867 官方实现网站&#xff1a;PRP 官方代码&#xff1a;https://github.…

Linux学习:初识Linux

目录 1. 引子&#xff1a;1.1 简述&#xff1a;操作系统1.2 学习工具 2. Linux操作系统中的一些基础概念与指令2.1 简单指令2.2 ls指令与文件2.3 cd指令与目录2.4 文件目录的新建与删除指令2.5 补充指令1&#xff1a;2.6 文件编辑与拷贝剪切2.7 文件的查看2.8 时间相关指令2.9 …

【代码】Python3|无GUI环境中使用Seaborn作图的学习路线及代码(阴影折线图)

我有个需求是需要画图&#xff0c;让GPT帮我生成了一下学习计划。 学习路线依照GPT的来的&#xff0c;使用的Prompt工具是https://github.com/JushBJJ/Mr.-Ranedeer-AI-Tutor。 文章目录 PrerequisiteMain Curriculum1.1 Seaborn介绍Seaborn基础保存图形为文件练习 1.2 单变量数…

day34贪心算法 part03

1005. K 次取反后最大化的数组和 简单 给你一个整数数组 nums 和一个整数 k &#xff0c;按以下方法修改该数组&#xff1a; 选择某个下标 i 并将 nums[i] 替换为 -nums[i] 。 重复这个过程恰好 k 次。可以多次选择同一个下标 i 。 以这种方式修改数组后&#xff0c;返回数…

【数据结构和算法初阶(C语言)】带环链表问题详解(快慢指针的烧脑应用)

目录 1.铺垫-----带环链表基本了解 2. 题目&#xff1a;环形链表 3.环形链表|| ​编辑 3.1题解1 3.2 题解2 4.总结 1.铺垫-----带环链表基本了解 环形链表题目启迪&#xff1a; 环形链表特点&#xff1a;遍历链表会出现一模一样的地址 2. 题目&#xff1a;环形链表 给…

C++输入输出(I\O)

我们知道C是由C语言发展而来的&#xff0c;几乎完全兼容C语言&#xff0c;换句话说&#xff0c;你可以在C里面编译C语言代码。如下图: C语言是面向过程的语言&#xff0c;C在C语言之上增加了面向对象以及泛型编程机制&#xff0c;因此C更适合中大型程序的开发&#xff0c;然而C…

Linux 设置快捷命令

以ll命令为例&#xff1a; 在 Linux 系统上&#xff0c;ll 命令通常不是一个独立的程序&#xff0c;而是 ls 命令的一个别名。 这个别名通常在用户的 shell 配置文件中定义&#xff0c;比如 .bashrc 或 .bash_aliases 文件中。 要在 Debian 上启用 ll 命令&#xff0c;你可以按…

李沐动手学习深度学习——4.5练习

1. 在本节的估计问题中使用λ的值进行实验。绘制训练和测试精度关于λ的函数。观察到了什么&#xff1f; 修改代码运行如图所示&#xff0c;可以发现对于lamda值的变化而言&#xff0c;对于训练loss和测试loss的影响不大。但是如果λ 太大后&#xff0c;train和test的loss会变得…

Cyber RT 组件

场景 无人车上的传感器数据可能需要被融合&#xff0c;比如在车辆上安装了多颗雷达&#xff0c;不同雷达由于安装位置与自身参数差异&#xff0c;可探测的角度、范围、距离等都是不尽相同的&#xff0c;现在需要将不同雷达感知到的数据融合在一起以建立车辆所处的完整环境&…

机器学习-面经(part5、KNN和SVM)

8. KNN 8.1 简述一下KNN算法的原理? 一句话概括:KNN的原理就是当预测一个新的值x的时候,根据它距离最近的K个点是什么类别来判断x属于哪个类别 工作原理:存在一个样本数据集合,也称作为训练样本集,并且样本集中每个数据都存在标签,即我们知道样本集中每一个数…

TypeError: the JSON object must be str, bytes or bytearray, not dict

参考文章&#xff1a;https://blog.csdn.net/yuan2019035055/article/details/124934362 Python基础系列&#xff08;一&#xff09;搞懂json数据解析与字典之间的关系 代码&#xff1a; 报错信息: TypeError: the JSON object must be str, bytes or bytearray, not dict …

局域网如何远程?

局域网远程一直是许多用户在处理远程连接需求时面临的一个难题。随着技术的不断进步&#xff0c;一种名为“天联”的组网解决方案应运而生。天联组网具有操作简单、跨平台应用、无网络要求以及独创的安全加速方案等独特优势&#xff0c;在解决各行业客户的远程连接需求方面发挥…

解决ipconfig不能使用的问题

问题所示&#xff1a;ipconfig不是内部或外部命令&#xff0c;也不是可运行的程序或批处理文件。 解决办法如下: 1.右击此电脑&#xff0c;点击属性设置&#xff1a; 2.点击高级系统设置 3.点击进入环境变量 4.在系统变量中进行设置&#xff0c;双击PATH进行配置 5.点击新建&am…

【如何在Docker中,修改已经挂载的卷(Volume)】

曾梦想执剑走天涯&#xff0c;我是程序猿【AK】 提示&#xff1a;添加投票&#xff01;&#xff01;&#xff01; 目录 简述概要知识图谱 简述概要 如何在Docker中&#xff0c;修改已经挂载的卷&#xff08;Volume&#xff09; 知识图谱 在Docker中&#xff0c;修改已经挂载…

matlab 提取分割位于多边形区域边缘内部或边缘上的点

[in,on] = inpolygon(xq,yq,xv,yv) xv 和 yv 为定义的多边形区域的,如xv = [1 4 4 1 1 ];yv = [1 1 4 4 1 ];注意最后一个数字与第一个重复,保证多边形闭合; xq 和 yq 为待查询的点in:在多边形内部和边缘的点序号on:仅在多边形边缘的点序号 提取分割方法: matrix=[xq yq…

智能汽车加速车规级存储应用DS2431P+TR 汽车级EEPROM 存储器IC

DS2431PT&R是一款1024位1-Wire EEPROM芯片&#xff0c;由四页存储区组成&#xff0c;每页256位。数据先被写入一个8字节暂存器中&#xff0c;经校验后复制到EEPROM存储器。该器件的特点是&#xff0c;四页存储区相互独立&#xff0c;可以单独进行写保护或进入EPROM仿真模式…

折线图实现柱状阴影背景的demo

这个是一个由官网的基础折线图实现的流程&#xff0c;将涉及到的知识点附上个人浅薄的见解&#xff0c;源码在最后&#xff0c;需要的可自取。 折线图 成果展示代码注解参数backgroundColordataZoomlegendtitlexAxisyAxisgridseries 源码 成果展示 官网的基础折线图&#xff…

【Python】OpenCV-使用ResNet50进行图像分类

使用ResNet50进行图像分类 如何使用ResNet50模型对图像进行分类。 import os import cv2 import numpy as np from tensorflow.keras.applications.resnet50 import ResNet50, preprocess_input, decode_predictions from tensorflow.keras.preprocessing import image# 设置…