mirror of
https://github.com/temporalio/skill-temporal-developer.git
synced 2026-09-14 13:52:58 +08:00
0c8586b4c2
* Add Java SDK reference files (11 files) Create complete Java reference documentation covering: - java.md: Entry point with quick start tutorial, key concepts - patterns.md: 17 patterns (signals, queries, updates, child workflows, saga, cancellation scopes, heartbeating, etc.) - determinism.md: Safe alternatives table, forbidden operations - determinism-protection.md: Convention-based enforcement (no sandbox) - error-handling.md: ApplicationFailure, retry/timeout config - gotchas.md: Non-deterministic operations, cancellation, heartbeating - testing.md: TestWorkflowEnvironment, Mockito mocking, replay testing - versioning.md: Workflow.getVersion(), worker versioning - data-handling.md: Jackson, PayloadConverter, encryption, search attributes - observability.md: SLF4J logging, Micrometer metrics - advanced-features.md: Schedules, async completion, worker tuning Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * Fix Java alignment issues from self-review - Reduce gotchas.md Non-Deterministic Operations from ~94 lines to ~12 (reference determinism.md instead of duplicating) - Remove Workflow Failure Exception Types duplication from error-handling.md (keep only in advanced-features.md) - Expand versioning.md Worker Versioning with Key Concepts, PINNED vs AUTO_UPGRADE, Deployment Strategies subsections - Fix section names to match Python reference style: Activity Heartbeat Details, Handling Activity Errors, Retry Policy Configuration, Workflow Test Environment, Mocking Activities, Workflow Replay Testing - Reduce data-handling.md Payload Encryption verbosity - Reduce observability.md Logger Customization verbosity - Reduce testing.md to single approach per section - Rename determinism.md "Convention-Based Enforcement" to "SDK Protection" - Fix handler guidance in patterns.md to match Python Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * Fix correctness issues in Java reference files - patterns.md: Fix Queries section — ActivityStub → typed interface (Workflow.newActivityStub returns the typed interface, not ActivityStub) - data-handling.md: Add missing ProtobufPayloadConverter to default converter chain (4th of 5 converters) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * Add Java to SKILL.md and core/determinism.md - SKILL.md: Add "Temporal Java" trigger phrase, update Overview to list Java, add Java entry to Getting Started references - core/determinism.md: Add Java entry to SDK Protection Mechanisms (no sandbox, convention-based, NonDeterministicException at replay) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * Apply manual editorial fixes to Java references - java.md: Remove "Understanding Replay" section (covered by Overview), simplify File Organization note (no sandbox rationale) - gotchas.md: Move Heartbeating before Cancellation, make Wrong Retry Classification brief with reference (not inline examples) - error-handling.md: Remove editorializing from Workflow Failure note - determinism-protection.md: Remove cross-language comparison paragraph (state Java's approach on its own terms) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * Add temporal-workflowcheck static analysis to Java determinism docs - determinism-protection.md: Add "Static Analysis with temporal-workflowcheck" section with Gradle/Maven setup, manual run, and suppression instructions. Beta warning included. - determinism.md: Update overview and SDK Protection to reference workflowcheck - core/determinism.md: Update Java entry in SDK Protection Mechanisms Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com> * Integrate feedback from Go PR into Java patterns - Updates: Add validator note — validators must not mutate state or block (matches note added to Python, TypeScript, Go, and core) - Saga Pattern: Use Workflow.newDetachedCancellationScope() for compensations so they execute even if the workflow is cancelled (mirrors Go's workflow.NewDisconnectedContext pattern) Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com> * docs: add @WorkflowInit description to java.md Key Concepts Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * mark java as supported * Apply suggestions from code review Co-authored-by: Brian Strauch <brian@brianstrauch.com> * strongly recommend java 21+ * Softened stance on static checker and replay testing. * address python/typescript sandboxing comment --------- Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com> Co-authored-by: Brian Strauch <brian.strauch@temporal.io> Co-authored-by: Brian Strauch <brian@brianstrauch.com>
189 lines
5.9 KiB
Markdown
189 lines
5.9 KiB
Markdown
# Java SDK Error Handling
|
|
|
|
## Overview
|
|
|
|
The Java SDK uses `ApplicationFailure` for application-specific errors and `RetryOptions` for retry configuration. Generally, the following information about errors and retryability applies across activities, child workflows and Nexus operations.
|
|
|
|
## Application Errors
|
|
|
|
```java
|
|
import io.temporal.activity.ActivityInterface;
|
|
import io.temporal.activity.ActivityMethod;
|
|
import io.temporal.failure.ApplicationFailure;
|
|
|
|
@ActivityInterface
|
|
public interface OrderActivities {
|
|
@ActivityMethod
|
|
void validateOrder(Order order);
|
|
}
|
|
|
|
public class OrderActivitiesImpl implements OrderActivities {
|
|
@Override
|
|
public void validateOrder(Order order) {
|
|
if (!order.isValid()) {
|
|
throw ApplicationFailure.newFailure(
|
|
"Invalid order",
|
|
"ValidationError"
|
|
);
|
|
}
|
|
}
|
|
}
|
|
```
|
|
|
|
Any exception that is not an `ApplicationFailure` is automatically converted to one, with the fully qualified class name as the type. For example, throwing `new NullPointerException("msg")` is equivalent to `ApplicationFailure.newFailure("msg", "java.lang.NullPointerException")`.
|
|
|
|
## Non-Retryable Errors
|
|
|
|
```java
|
|
import io.temporal.failure.ApplicationFailure;
|
|
|
|
public class PaymentActivitiesImpl implements PaymentActivities {
|
|
@Override
|
|
public String chargeCard(String cardNumber, double amount) {
|
|
if (!isValidCard(cardNumber)) {
|
|
throw ApplicationFailure.newNonRetryableFailure(
|
|
"Permanent failure - invalid credit card",
|
|
"PaymentError"
|
|
);
|
|
}
|
|
return processPayment(cardNumber, amount);
|
|
}
|
|
}
|
|
```
|
|
|
|
You can also mark error types as non-retryable via `RetryOptions.setDoNotRetry()`:
|
|
|
|
```java
|
|
RetryOptions retryOptions = RetryOptions.newBuilder()
|
|
.setDoNotRetry(
|
|
CreditCardProcessingException.class.getName(),
|
|
"ValidationError"
|
|
)
|
|
.build();
|
|
```
|
|
|
|
Use `newNonRetryableFailure()` when the **activity implementer** knows the error is permanent. Use `setDoNotRetry()` when the **caller** wants to control retryability.
|
|
|
|
## Activity Errors
|
|
|
|
Activity failures are always wrapped in `ActivityFailure`. The original exception becomes the `cause`:
|
|
|
|
- `ActivityFailure` → `ApplicationFailure` (application error)
|
|
- `ActivityFailure` → `TimeoutFailure` (timeout)
|
|
- `ActivityFailure` → `CanceledFailure` (cancellation)
|
|
|
|
## Handling Activity Errors
|
|
|
|
```java
|
|
import io.temporal.failure.ActivityFailure;
|
|
import io.temporal.failure.ApplicationFailure;
|
|
import io.temporal.failure.TimeoutFailure;
|
|
import io.temporal.workflow.Workflow;
|
|
|
|
public class MyWorkflowImpl implements MyWorkflow {
|
|
@Override
|
|
public String run() {
|
|
try {
|
|
return activities.riskyOperation();
|
|
} catch (ActivityFailure af) {
|
|
if (af.getCause() instanceof ApplicationFailure) {
|
|
ApplicationFailure appFailure = (ApplicationFailure) af.getCause();
|
|
String type = appFailure.getType();
|
|
// Handle based on error type
|
|
} else if (af.getCause() instanceof TimeoutFailure) {
|
|
// Handle timeout
|
|
}
|
|
throw ApplicationFailure.newFailure(
|
|
"Workflow failed due to activity error",
|
|
"WorkflowError"
|
|
);
|
|
}
|
|
}
|
|
}
|
|
```
|
|
|
|
## Retry Policy Configuration
|
|
|
|
```java
|
|
import io.temporal.activity.ActivityOptions;
|
|
import io.temporal.common.RetryOptions;
|
|
import io.temporal.workflow.Workflow;
|
|
|
|
import java.time.Duration;
|
|
|
|
public class MyWorkflowImpl implements MyWorkflow {
|
|
|
|
private final MyActivities activities = Workflow.newActivityStub(
|
|
MyActivities.class,
|
|
ActivityOptions.newBuilder()
|
|
.setStartToCloseTimeout(Duration.ofMinutes(10))
|
|
.setRetryOptions(RetryOptions.newBuilder()
|
|
.setMaximumInterval(Duration.ofMinutes(1))
|
|
.setMaximumAttempts(5)
|
|
.setDoNotRetry("ValidationError", "PaymentError")
|
|
.build())
|
|
.build()
|
|
);
|
|
|
|
@Override
|
|
public String run() {
|
|
return activities.myActivity();
|
|
}
|
|
}
|
|
```
|
|
|
|
Only set options such as `maximumInterval`, `maximumAttempts` etc. if you have a domain-specific reason to. If not, prefer to leave them at their defaults.
|
|
|
|
## Timeout Configuration
|
|
|
|
```java
|
|
ActivityOptions options = ActivityOptions.newBuilder()
|
|
.setStartToCloseTimeout(Duration.ofMinutes(5)) // Single attempt
|
|
.setScheduleToCloseTimeout(Duration.ofMinutes(30)) // Including retries
|
|
.setHeartbeatTimeout(Duration.ofMinutes(2)) // Between heartbeats
|
|
.build();
|
|
```
|
|
|
|
## Workflow Failure
|
|
|
|
**IMPORTANT:** Only `ApplicationFailure` causes a workflow to fail. Any other exception thrown from workflow code causes the workflow task to retry indefinitely, not the workflow itself.
|
|
|
|
```java
|
|
import io.temporal.failure.ApplicationFailure;
|
|
|
|
public class MyWorkflowImpl implements MyWorkflow {
|
|
@Override
|
|
public String run() {
|
|
if (someCondition) {
|
|
throw ApplicationFailure.newFailure(
|
|
"Cannot process order",
|
|
"BusinessError"
|
|
);
|
|
}
|
|
return "success";
|
|
}
|
|
}
|
|
```
|
|
|
|
To allow other exception types to fail the workflow instead of causing infinite task retries, see `references/java/advanced-features.md` for configuring `setFailWorkflowExceptionTypes()`.
|
|
|
|
Use checked exceptions with `Workflow.wrap()` to rethrow them as unchecked:
|
|
|
|
```java
|
|
try {
|
|
return someCall();
|
|
} catch (Exception e) {
|
|
throw Workflow.wrap(e);
|
|
}
|
|
```
|
|
|
|
## Best Practices
|
|
|
|
1. Use specific error types for different failure modes
|
|
2. Mark permanent failures as non-retryable
|
|
3. Configure appropriate retry policies
|
|
4. Log errors before re-raising
|
|
5. Catch `ActivityFailure` (not `ApplicationFailure`) for activity failures in workflows
|
|
6. Design code to be idempotent for safe retries (see more at `references/core/patterns.md`)
|
|
7. Use `ApplicationFailure.newFailure()` to fail workflows — other exceptions cause infinite task retries
|