Fix test server and tests for Temporal Server 1.31.0 compatibility

- Update test server nexus failure wrapping to match real server behavior:
  remove ApplicationFailureInfo wrapping for non-temporal failures and
  remove redundant message from handler error top-level failure
- Fix NexusWorkflowTest assertions for updated failure structure
- Fix testNexusOperationTimeout_AfterCancel to handle intermediate WFTs
  caused by NEXUS_OPERATION_CANCEL_REQUEST_FAILED events
- Fix testNexusOperationScheduleToStartTimeout timing for real server
- Use unique workflow IDs in WorkflowIdConflictPolicyTest for real server
- Add component.callbacks.allowedAddresses config to CI server startup

Author: bergundy

.github/workflows/ci.yml

@@ -94,7 +94,8 @@ jobs:
             --search-attribute CustomBoolField=Bool \
             --dynamic-config-value system.enableActivityEagerExecution=true \
             --dynamic-config-value frontend.workerVersioningDataAPIs=true \
-            --dynamic-config-value history.enableRequestIdRefLinks=true &
+            --dynamic-config-value history.enableRequestIdRefLinks=true \
+            --dynamic-config-value 'component.callbacks.allowedAddresses=[{"Pattern":"localhost:7243","AllowInsecure":true}]' &
         sleep 10s
 
     - name: Run unit tests

temporal-test-server/src/main/java/io/temporal/internal/testservice/TestWorkflowService.java

@@ -1171,7 +1171,6 @@ public void completeNexusOperation(
 
   private static Failure handlerErrorToFailure(HandlerError err) {
     return Failure.newBuilder()
-        .setMessage(err.getFailure().getMessage())
         .setNexusHandlerFailureInfo(
             NexusHandlerFailureInfo.newBuilder()
                 .setType(err.getErrorType())
@@ -1196,13 +1195,6 @@ private static Failure nexusFailureToAPIFailure(
       } catch (InvalidProtocolBufferException e) {
         throw new RuntimeException(e);
       }
-    } else {
-      Payloads payloads = nexusFailureMetadataToPayloads(failure);
-      ApplicationFailureInfo.Builder applicationFailureInfo = ApplicationFailureInfo.newBuilder();
-      applicationFailureInfo.setType("NexusFailure");
-      applicationFailureInfo.setDetails(payloads);
-      applicationFailureInfo.setNonRetryable(!retryable);
-      apiFailure.setApplicationFailureInfo(applicationFailureInfo.build());
     }
     apiFailure.setMessage(failure.getMessage());
     return apiFailure.build();

temporal-test-server/src/test/java/io/temporal/testserver/functional/NexusWorkflowTest.java

@@ -628,7 +628,7 @@ public void testNexusOperationTimeout_AfterStart() {
     }
   }
 
-  @Test(timeout = 30000)
+  @Test(timeout = 60000)
   public void testNexusOperationTimeout_AfterCancel() {
     String operationId = UUID.randomUUID().toString();
     CompletableFuture<?> nexusPoller =
@@ -644,7 +644,7 @@ public void testNexusOperationTimeout_AfterCancel() {
           pollResp.getTaskToken(),
           newScheduleOperationCommand(
               defaultScheduleOperationAttributes()
-                  .setScheduleToCloseTimeout(Durations.fromSeconds(22))));
+                  .setScheduleToCloseTimeout(Durations.fromSeconds(10))));
       testWorkflowRule.assertHistoryEvent(
           execution.getWorkflowId(), EventType.EVENT_TYPE_NEXUS_OPERATION_SCHEDULED);
 
@@ -669,24 +669,21 @@ public void testNexusOperationTimeout_AfterCancel() {
               .build();
       completeWorkflowTask(pollResp.getTaskToken(), cancelCmd);
 
-      // Poll for cancellation task but do not complete it
-      PollNexusTaskQueueResponse nexusPollResp = pollNexusTask().get();
-      Assert.assertTrue(nexusPollResp.getRequest().hasCancelOperation());
-
-      // Request timeout and long poll deadline are both 10s, so sleep to give some buffer so poll
-      // request doesn't timeout.
-      Thread.sleep(2000);
-
-      // Poll for cancellation task again
-      nexusPollResp = pollNexusTask().get();
-      Assert.assertTrue(nexusPollResp.getRequest().hasCancelOperation());
-
-      // Request timeout and long poll deadline are both 10s, so sleep to give some buffer so poll
-      // request doesn't timeout.
-      Thread.sleep(2000);
+      // The cancel task may not be completed in time, causing a
+      // NEXUS_OPERATION_CANCEL_REQUEST_FAILED event and a new WFT. Handle any intermediate WFTs
+      // until the operation actually times out.
+      while (true) {
+        pollResp = pollWorkflowTask();
+        events =
+            testWorkflowRule.getHistoryEvents(
+                execution.getWorkflowId(), EventType.EVENT_TYPE_NEXUS_OPERATION_TIMED_OUT);
+        if (!events.isEmpty()) {
+          break;
+        }
+        // Not timed out yet, acknowledge the WFT and keep waiting
+        completeWorkflowTask(pollResp.getTaskToken());
+      }
 
-      // Poll to wait for new task after operation times out
-      pollResp = pollWorkflowTask();
       completeWorkflow(pollResp.getTaskToken());
 
       events =
@@ -710,18 +707,21 @@ public void testNexusOperationTimeout_AfterCancel() {
     }
   }
 
-  @Test
+  @Test(timeout = 30000)
   public void testNexusOperationScheduleToStartTimeout() {
     WorkflowStub stub = newWorkflowStub("TestNexusOperationScheduleToStartTimeoutWorkflow");
     WorkflowExecution execution = stub.start();
 
+    // Use a longer schedule-to-start timeout so the nexus poll has time to return on real server
+    int scheduleToStartSeconds = testWorkflowRule.isUseExternalService() ? 10 : 1;
+
     // Get first WFT and respond with ScheduleNexusOperation command with schedule-to-start timeout
     PollWorkflowTaskQueueResponse pollResp = pollWorkflowTask();
     completeWorkflowTask(
         pollResp.getTaskToken(),
         newScheduleOperationCommand(
             defaultScheduleOperationAttributes()
-                .setScheduleToStartTimeout(Durations.fromSeconds(1))
+                .setScheduleToStartTimeout(Durations.fromSeconds(scheduleToStartSeconds))
                 .setScheduleToCloseTimeout(Durations.fromSeconds(30))));
     testWorkflowRule.assertHistoryEvent(
         execution.getWorkflowId(), EventType.EVENT_TYPE_NEXUS_OPERATION_SCHEDULED);
@@ -746,7 +746,7 @@ public void testNexusOperationScheduleToStartTimeout() {
       Assert.assertTrue("OPERATION_TIMEOUT should be positive", operationTimeoutMs > 0);
 
       // Sleep longer than schedule-to-start timeout to trigger the timeout
-      testWorkflowRule.sleep(Duration.ofSeconds(2));
+      testWorkflowRule.sleep(Duration.ofSeconds(scheduleToStartSeconds + 1));
     } catch (Exception e) {
       Assert.fail(e.getMessage());
     }
@@ -885,8 +885,6 @@ public void testNexusOperationError() {
       Assert.assertEquals("nexus operation completed unsuccessfully", failure.getMessage());
       io.temporal.api.failure.v1.Failure cause = failure.getCause();
       Assert.assertEquals("deliberate test failure", cause.getMessage());
-      Assert.assertTrue(cause.hasApplicationFailureInfo());
-      Assert.assertEquals("NexusFailure", cause.getApplicationFailureInfo().getType());
     } catch (Exception e) {
       Assert.fail(e.getMessage());
     } finally {
@@ -947,9 +945,9 @@ public void testNexusOperationHandlerError() {
       assertOperationFailureInfo(failure.getNexusOperationExecutionFailureInfo());
       Assert.assertEquals("nexus operation completed unsuccessfully", failure.getMessage());
       io.temporal.api.failure.v1.Failure cause = failure.getCause();
-      Assert.assertTrue(cause.getMessage().endsWith("deliberate terminal error"));
       Assert.assertTrue(cause.hasNexusHandlerFailureInfo());
       Assert.assertEquals("BAD_REQUEST", cause.getNexusHandlerFailureInfo().getType());
+      Assert.assertEquals("deliberate terminal error", cause.getCause().getMessage());
     } catch (Exception e) {
       Assert.fail(e.getMessage());
     } finally {
@@ -1016,9 +1014,9 @@ public void testNexusOperationHandlerTemporalFailure() {
       assertOperationFailureInfo(failure.getNexusOperationExecutionFailureInfo());
       Assert.assertEquals("nexus operation completed unsuccessfully", failure.getMessage());
       io.temporal.api.failure.v1.Failure cause = failure.getCause();
-      Assert.assertTrue(cause.getMessage().endsWith("deliberate terminal error"));
       Assert.assertTrue(cause.hasNexusHandlerFailureInfo());
       Assert.assertEquals("BAD_REQUEST", cause.getNexusHandlerFailureInfo().getType());
+      Assert.assertEquals("deliberate terminal error", cause.getCause().getMessage());
     } catch (Exception e) {
       Assert.fail(e.getMessage());
     } finally {

temporal-test-server/src/test/java/io/temporal/testserver/functional/WorkflowIdConflictPolicyTest.java

@@ -41,7 +41,7 @@ public class WorkflowIdConflictPolicyTest {
 
   @Test
   public void conflictPolicyUseExisting() {
-    String workflowId = "conflict-policy-use-existing";
+    String workflowId = "conflict-policy-use-existing-" + randomUUID();
     String requestId = randomUUID().toString();
 
     // Start workflow
@@ -232,7 +232,7 @@ public void conflictPolicyUseExisting() {
 
   @Test
   public void conflictPolicyFail() {
-    String workflowId = "conflict-policy-fail";
+    String workflowId = "conflict-policy-fail-" + randomUUID();
     WorkflowOptions options =
         WorkflowOptions.newBuilder()
             .setWorkflowId(workflowId)