Skip to content

[ISSUE #10766] Fail fast on telemetry dispatch failure #10767

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
Aias00 wants to merge 2 commits into
base: develop
Choose a base branch
from
+102 −7
Open

[ISSUE #10766] Fail fast on telemetry dispatch failure #10767

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
52603a9
[ISSUE #10766] Fail fast on telemetry dispatch failure
Aias00 Aug 3, 2026
db27454
fix(proxy): preserve telemetry write API compatibility
Aias00 Aug 4, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
36 changes: 30 additions & 6 deletions proxy/src/main/java/org/apache/rocketmq/proxy/grpc/v2/channel/GrpcClientChannel.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -51,6 +51,7 @@
import org.apache.rocketmq.proxy.service.relay.ProxyRelayResult;
import org.apache.rocketmq.proxy.service.relay.ProxyRelayService;
import org.apache.rocketmq.proxy.service.transaction.TransactionData;
import org.apache.rocketmq.remoting.protocol.ResponseCode;
import org.apache.rocketmq.remoting.protocol.RemotingCommand;
import org.apache.rocketmq.remoting.protocol.body.ConsumeMessageDirectlyResult;
import org.apache.rocketmq.remoting.protocol.body.ConsumerRunningInfo;
Expand Down Expand Up @@ -235,24 +236,32 @@ protected CompletableFuture<Void> processGetConsumerRunningInfo(RemotingCommand
if (Objects.isNull(header) || !header.isJstackEnable()) {
return CompletableFuture.completedFuture(null);
}
this.writeTelemetryCommand(TelemetryCommand.newBuilder()
String nonce = this.grpcChannelManager.addResponseFuture(responseFuture);
boolean written = this.tryWriteTelemetryCommand(TelemetryCommand.newBuilder()
.setPrintThreadStackTraceCommand(PrintThreadStackTraceCommand.newBuilder()
.setNonce(this.grpcChannelManager.addResponseFuture(responseFuture))
.setNonce(nonce)
.build())
.build());
if (!written) {
this.completeResponseFutureOnWriteFailure(nonce, responseFuture);
}
return CompletableFuture.completedFuture(null);
}

@Override
protected CompletableFuture<Void> processConsumeMessageDirectly(RemotingCommand command,
ConsumeMessageDirectlyResultRequestHeader header,
MessageExt messageExt, CompletableFuture<ProxyRelayResult<ConsumeMessageDirectlyResult>> responseFuture) {
this.writeTelemetryCommand(TelemetryCommand.newBuilder()
String nonce = this.grpcChannelManager.addResponseFuture(responseFuture);
boolean written = this.tryWriteTelemetryCommand(TelemetryCommand.newBuilder()
.setVerifyMessageCommand(VerifyMessageCommand.newBuilder()
.setNonce(this.grpcChannelManager.addResponseFuture(responseFuture))
.setNonce(nonce)
.setMessage(GrpcConverter.getInstance().buildMessage(messageExt))
.build())
.build());
if (!written) {
this.completeResponseFutureOnWriteFailure(nonce, responseFuture);
}
return CompletableFuture.completedFuture(null);
}

Expand All @@ -261,26 +270,41 @@ public String getClientId() {
}

public void writeTelemetryCommand(TelemetryCommand command) {
this.tryWriteTelemetryCommand(command);
}

private boolean tryWriteTelemetryCommand(TelemetryCommand command) {
StreamObserver<TelemetryCommand> observer = this.telemetryCommandRef.get();
if (observer == null) {
log.warn("telemetry command observer is null when try to write data. command:{}, channel:{}", TextFormat.shortDebugString(command), this);
return;
return false;
}
synchronized (this.telemetryWriteLock) {
observer = this.telemetryCommandRef.get();
if (observer == null) {
log.warn("telemetry command observer is null when try to write data. command:{}, channel:{}", TextFormat.shortDebugString(command), this);
return;
return false;
}
try {
observer.onNext(command);
return true;
} catch (StatusRuntimeException | IllegalStateException exception) {
log.warn("write telemetry failed. command:{}", command, exception);
this.clearClientObserver(observer);
return false;
}
}
}

private <T> void completeResponseFutureOnWriteFailure(String nonce,
CompletableFuture<ProxyRelayResult<T>> responseFuture) {
CompletableFuture<ProxyRelayResult<T>> registeredFuture =
this.grpcChannelManager.getAndRemoveResponseFuture(nonce);
CompletableFuture<ProxyRelayResult<T>> future =
registeredFuture == null ? responseFuture : registeredFuture;
future.complete(new ProxyRelayResult<>(ResponseCode.SYSTEM_BUSY, "write telemetry command failed", null));
}

@Override
public String toString() {
return MoreObjects.toStringHelper(this)
Expand Down
73 changes: 72 additions & 1 deletion proxy/src/test/java/org/apache/rocketmq/proxy/grpc/v2/channel/GrpcClientChannelTest.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,24 +20,41 @@
import apache.rocketmq.v2.Publishing;
import apache.rocketmq.v2.Resource;
import apache.rocketmq.v2.Settings;
import apache.rocketmq.v2.TelemetryCommand;
import io.grpc.stub.StreamObserver;
import java.nio.charset.StandardCharsets;
import java.util.concurrent.CompletableFuture;
import org.apache.commons.lang3.RandomStringUtils;
import org.apache.rocketmq.common.message.MessageExt;
import org.apache.rocketmq.proxy.common.ProxyContext;
import org.apache.rocketmq.proxy.config.InitConfigTest;
import org.apache.rocketmq.proxy.grpc.v2.common.GrpcClientSettingsManager;
import org.apache.rocketmq.proxy.processor.channel.ChannelProtocolType;
import org.apache.rocketmq.proxy.processor.channel.RemoteChannel;
import org.apache.rocketmq.proxy.remoting.channel.RemotingChannel;
import org.apache.rocketmq.proxy.service.relay.ProxyRelayResult;
import org.apache.rocketmq.proxy.service.relay.ProxyRelayService;
import org.apache.rocketmq.remoting.protocol.RemotingCommand;
import org.apache.rocketmq.remoting.protocol.ResponseCode;
import org.apache.rocketmq.remoting.protocol.body.ConsumeMessageDirectlyResult;
import org.apache.rocketmq.remoting.protocol.body.ConsumerRunningInfo;
import org.apache.rocketmq.remoting.protocol.header.ConsumeMessageDirectlyResultRequestHeader;
import org.apache.rocketmq.remoting.protocol.header.GetConsumerRunningInfoRequestHeader;
import org.junit.Before;
import org.junit.Test;
import org.junit.runner.RunWith;
import org.mockito.Mock;
import org.mockito.junit.MockitoJUnitRunner;

import static org.junit.Assert.assertEquals;
import static org.junit.Assert.assertFalse;
import static org.junit.Assert.assertNull;
import static org.junit.Assert.assertTrue;
import static org.mockito.ArgumentMatchers.any;
import static org.mockito.ArgumentMatchers.eq;
import static org.mockito.Mockito.doThrow;
import static org.mockito.Mockito.mock;
import static org.mockito.Mockito.verify;
import static org.mockito.Mockito.when;

@RunWith(MockitoJUnitRunner.class)
Expand Down Expand Up @@ -79,4 +96,58 @@ public void testChannelExtendAttributeParse() {
assertEquals(clientSettings, GrpcClientChannel.parseChannelExtendAttribute(this.grpcClientChannel));
assertNull(GrpcClientChannel.parseChannelExtendAttribute(mock(RemotingChannel.class)));
}
}

@Test
public void testGetConsumerRunningInfoShouldFailFastWhenObserverIsMissing() throws Exception {
CompletableFuture<ProxyRelayResult<ConsumerRunningInfo>> responseFuture = new CompletableFuture<>();
when(grpcChannelManager.addResponseFuture(eq(responseFuture))).thenReturn("nonce-1");
when(grpcChannelManager.<ConsumerRunningInfo>getAndRemoveResponseFuture(eq("nonce-1"))).thenReturn(responseFuture);

GetConsumerRunningInfoRequestHeader header = new GetConsumerRunningInfoRequestHeader();
header.setJstackEnable(true);

grpcClientChannel.processGetConsumerRunningInfo(mock(RemotingCommand.class), header, responseFuture).get();

assertTrue(responseFuture.isDone());
ProxyRelayResult<ConsumerRunningInfo> result = responseFuture.get();
assertEquals(ResponseCode.SYSTEM_BUSY, result.getCode());
assertEquals("write telemetry command failed", result.getRemark());
verify(grpcChannelManager).getAndRemoveResponseFuture("nonce-1");
}

@Test
public void testConsumeMessageDirectlyShouldFailFastWhenObserverWriteFails() throws Exception {
StreamObserver<TelemetryCommand> observer = mock(StreamObserver.class);
doThrow(new IllegalStateException("stream closed")).when(observer).onNext(any(TelemetryCommand.class));
grpcClientChannel.setClientObserver(observer);

CompletableFuture<ProxyRelayResult<ConsumeMessageDirectlyResult>> responseFuture = new CompletableFuture<>();
when(grpcChannelManager.addResponseFuture(eq(responseFuture))).thenReturn("nonce-2");
when(grpcChannelManager.<ConsumeMessageDirectlyResult>getAndRemoveResponseFuture(eq("nonce-2"))).thenReturn(responseFuture);

grpcClientChannel.processConsumeMessageDirectly(
mock(RemotingCommand.class),
new ConsumeMessageDirectlyResultRequestHeader(),
buildMessageExt(),
responseFuture
).get();

assertTrue(responseFuture.isDone());
ProxyRelayResult<ConsumeMessageDirectlyResult> result = responseFuture.get();
assertEquals(ResponseCode.SYSTEM_BUSY, result.getCode());
assertEquals("write telemetry command failed", result.getRemark());
verify(grpcChannelManager).getAndRemoveResponseFuture("nonce-2");
assertFalse(grpcClientChannel.isOpen());
}

private MessageExt buildMessageExt() {
MessageExt messageExt = new MessageExt();
messageExt.setTopic("test-topic");
messageExt.setBody("hello".getBytes(StandardCharsets.UTF_8));
messageExt.setMsgId("msg-id");
messageExt.setBornTimestamp(System.currentTimeMillis());
messageExt.setStoreTimestamp(System.currentTimeMillis());
messageExt.putUserProperty("test", "true");
return messageExt;
}
}
Loading