Skip to content

Retry transient errors in ActivityCompletionClient.heartbeat #2985

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
arimu1 wants to merge 2 commits into
base: main
Choose a base branch
from
Open

Retry transient errors in ActivityCompletionClient.heartbeat #2985

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
144d991
Retry transient gRPC errors in ManualActivityCompletionClient.recordH…
arimu1 Jul 31, 2026
a7ff2b4
Merge remote-tracking branch 'upstream/main' into fix/2984-heartbeat-…
arimu1 Oct 31, 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
12 changes: 10 additions & 2 deletions ...rc/main/java/io/temporal/internal/client/external/ManualActivityCompletionClientImpl.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -210,7 +210,9 @@ public void recordHeartbeat(@Nullable Object details) throws CanceledFailure {
payloads.ifPresent(builder::setDetails);
RecordActivityTaskHeartbeatRequest request = storeOutbound(builder.build());
RecordActivityTaskHeartbeatResponse status =
ActivityClientHelper.sendHeartbeatRequest(service, request, metricsScope);
grpcRetryer.retryWithResult(
() -> ActivityClientHelper.sendHeartbeatRequest(service, request, metricsScope),
replyGrpcRetryerOptions);
if (status.getCancelRequested()) {
throw new ActivityCanceledException();
} else if (status.getActivityReset()) {
Expand All @@ -229,7 +231,11 @@ public void recordHeartbeat(@Nullable Object details) throws CanceledFailure {
payloads.ifPresent(builder::setDetails);
RecordActivityTaskHeartbeatByIdRequest request = storeOutbound(builder.build());
RecordActivityTaskHeartbeatByIdResponse status =
ActivityClientHelper.recordActivityTaskHeartbeatById(service, request, metricsScope);
grpcRetryer.retryWithResult(
() ->
ActivityClientHelper.recordActivityTaskHeartbeatById(
service, request, metricsScope),
replyGrpcRetryerOptions);
if (status.getCancelRequested()) {
throw new ActivityCanceledException();
} else if (status.getActivityReset()) {
Expand All @@ -238,6 +244,8 @@ public void recordHeartbeat(@Nullable Object details) throws CanceledFailure {
throw new ActivityPausedException();
}
}
} catch (ActivityCompletionException e) {
throw e;
} catch (Exception e) {
processException(e);
}
Expand Down
153 changes: 153 additions & 0 deletions ...est/java/io/temporal/internal/client/external/ManualActivityCompletionClientImplTest.java
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,18 +1,32 @@
package io.temporal.internal.client.external;

import static org.junit.Assert.assertEquals;
import static org.junit.Assert.assertSame;
import static org.junit.Assert.assertThrows;
import static org.junit.Assert.assertTrue;
import static org.mockito.ArgumentMatchers.any;
import static org.mockito.Mockito.mock;
import static org.mockito.Mockito.never;
import static org.mockito.Mockito.times;
import static org.mockito.Mockito.verify;
import static org.mockito.Mockito.when;

import com.uber.m3.tally.NoopScope;
import io.grpc.Status;
import io.grpc.StatusRuntimeException;
import io.temporal.api.common.v1.Payload;
import io.temporal.api.common.v1.WorkflowExecution;
import io.temporal.api.workflowservice.v1.GetSystemInfoResponse;
import io.temporal.api.workflowservice.v1.RecordActivityTaskHeartbeatByIdRequest;
import io.temporal.api.workflowservice.v1.RecordActivityTaskHeartbeatByIdResponse;
import io.temporal.api.workflowservice.v1.RecordActivityTaskHeartbeatRequest;
import io.temporal.api.workflowservice.v1.RecordActivityTaskHeartbeatResponse;
import io.temporal.api.workflowservice.v1.WorkflowServiceGrpc;
import io.temporal.client.ActivityCanceledException;
import io.temporal.client.ActivityCompletionFailureException;
import io.temporal.client.ActivityNotExistsException;
import io.temporal.common.converter.DefaultDataConverter;
import io.temporal.common.converter.GlobalDataConverter;
import io.temporal.failure.ApplicationFailure;
import io.temporal.internal.payload.storage.ExternalStorageRunner;
import io.temporal.payload.storage.ExternalStorage;
Expand All @@ -21,16 +35,25 @@
import io.temporal.payload.storage.StorageDriverClaim;
import io.temporal.payload.storage.StorageDriverRetrieveContext;
import io.temporal.payload.storage.StorageDriverStoreContext;
import io.temporal.serviceclient.RpcRetryOptions;
import io.temporal.serviceclient.WorkflowServiceStubs;
import io.temporal.serviceclient.WorkflowServiceStubsOptions;
import java.time.Duration;
import java.util.List;
import java.util.concurrent.CompletableFuture;
import java.util.concurrent.atomic.AtomicInteger;
import org.junit.Before;
import org.junit.Test;

public class ManualActivityCompletionClientImplTest {
private static final byte[] TASK_TOKEN = new byte[] {1, 2, 3};
private static final String NAMESPACE = "test-namespace";
private static final String IDENTITY = "test-identity";
private static final String ACTIVITY_ID = "activity-1";

private final RuntimeException storageFailure = new RuntimeException("storage failed");
private WorkflowServiceStubs service;
private WorkflowServiceGrpc.WorkflowServiceBlockingStub blockingStub;
private ExternalStorageRunner externalStorage;

@Before
Expand Down Expand Up @@ -86,6 +109,136 @@ public void byIdCancellationIgnoresStorageFailure() {
verify(service, never()).blockingStub();
}

@Test
public void recordHeartbeatRetriesTransientErrorsThenSucceeds() {
AtomicInteger attempts = new AtomicInteger();
prepareHeartbeatStub();
when(blockingStub.recordActivityTaskHeartbeat(any(RecordActivityTaskHeartbeatRequest.class)))
.thenAnswer(
invocation -> {
if (attempts.getAndIncrement() < 2) {
throw new StatusRuntimeException(Status.RESOURCE_EXHAUSTED);
}
return RecordActivityTaskHeartbeatResponse.getDefaultInstance();
});

newClientWithTaskToken().recordHeartbeat("details");

assertEquals(3, attempts.get());
verify(blockingStub, times(3)).recordActivityTaskHeartbeat(any());
}

@Test
public void recordHeartbeatByIdRetriesTransientErrorsThenSucceeds() {
AtomicInteger attempts = new AtomicInteger();
prepareHeartbeatStub();
when(blockingStub.recordActivityTaskHeartbeatById(
any(RecordActivityTaskHeartbeatByIdRequest.class)))
.thenAnswer(
invocation -> {
if (attempts.getAndIncrement() < 2) {
throw new StatusRuntimeException(Status.UNAVAILABLE);
}
return RecordActivityTaskHeartbeatByIdResponse.getDefaultInstance();
});

newClientWithActivityId().recordHeartbeat("details");

assertEquals(3, attempts.get());
verify(blockingStub, times(3)).recordActivityTaskHeartbeatById(any());
}

@Test
public void recordHeartbeatDoesNotRetryNotFound() {
prepareHeartbeatStub();
when(blockingStub.recordActivityTaskHeartbeat(any(RecordActivityTaskHeartbeatRequest.class)))
.thenThrow(new StatusRuntimeException(Status.NOT_FOUND));

assertThrows(
ActivityNotExistsException.class,
() -> newClientWithTaskToken().recordHeartbeat("details"));
verify(blockingStub, times(1)).recordActivityTaskHeartbeat(any());
}

@Test
public void recordHeartbeatThrowsCanceledWhenServerRequestsCancel() {
prepareHeartbeatStub();
when(blockingStub.recordActivityTaskHeartbeat(any(RecordActivityTaskHeartbeatRequest.class)))
.thenReturn(
RecordActivityTaskHeartbeatResponse.newBuilder().setCancelRequested(true).build());

assertThrows(
ActivityCanceledException.class, () -> newClientWithTaskToken().recordHeartbeat("details"));
}

@Test
public void recordHeartbeatFailsAfterExhaustingRetries() {
prepareHeartbeatStub();
when(blockingStub.recordActivityTaskHeartbeat(any(RecordActivityTaskHeartbeatRequest.class)))
.thenThrow(new StatusRuntimeException(Status.RESOURCE_EXHAUSTED));

ActivityCompletionFailureException failure =
assertThrows(
ActivityCompletionFailureException.class,
() -> newClientWithTaskToken().recordHeartbeat("details"));
assertTrue(failure.getCause() instanceof StatusRuntimeException);
assertEquals(
Status.Code.RESOURCE_EXHAUSTED,
((StatusRuntimeException) failure.getCause()).getStatus().getCode());
verify(blockingStub, times(5)).recordActivityTaskHeartbeat(any());
}

private void prepareHeartbeatStub() {
blockingStub = mock(WorkflowServiceGrpc.WorkflowServiceBlockingStub.class);
when(service.blockingStub()).thenReturn(blockingStub);
when(blockingStub.withOption(any(), any())).thenReturn(blockingStub);
when(service.getOptions())
.thenReturn(
WorkflowServiceStubsOptions.newBuilder()
.setRpcRetryOptions(
RpcRetryOptions.newBuilder()
.setInitialInterval(Duration.ofMillis(1))
.setCongestionInitialInterval(Duration.ofMillis(1))
.setMaximumInterval(Duration.ofMillis(10))
.setExpiration(Duration.ofSeconds(2))
.setMaximumAttempts(5)
.setBackoffCoefficient(1.0)
.validateBuildWithDefaults())
.validateAndBuildWithDefaults());
}

private ManualActivityCompletionClientImpl newClientWithTaskToken() {
return new ManualActivityCompletionClientImpl(
service,
NAMESPACE,
IDENTITY,
GlobalDataConverter.get(),
new NoopScope(),
TASK_TOKEN,
null,
null,
null,
null,
null);
}

private ManualActivityCompletionClientImpl newClientWithActivityId() {
WorkflowExecution execution =
WorkflowExecution.newBuilder().setWorkflowId("wf-id").setRunId("run-id").build();
return new ManualActivityCompletionClientImpl(
service,
NAMESPACE,
IDENTITY,
GlobalDataConverter.get(),
new NoopScope(),
null,
execution,
ACTIVITY_ID,
null,
null,
null);
}

private ManualActivityCompletionClientImpl taskTokenClient() {
return new ManualActivityCompletionClientImpl(
service,
Expand Down
Loading