chore: Rerun flaky tests on CI if they failed

.github/workflows/build-and-test.yml

@@ -30,7 +30,7 @@ jobs:
         distribution: 'temurin'
         cache: maven
     - name: Build with Maven and run tests
-      run: mvn -B package --file pom.xml -fae ${{ matrix.os == 'windows-latest' && '-DskipDockerTests=true' || '' }}
+      run: mvn -B install --file pom.xml -fae ${{ matrix.os == 'windows-latest' && '-DskipDockerTests=true' || '' }}
     - name: Upload Test Reports
       if: failure()
       uses: actions/upload-artifact@v7

pom.xml

@@ -662,6 +662,43 @@
                 <quarkus.jvm.args></quarkus.jvm.args>
             </properties>
         </profile>
+        <profile>
+            <id>flaky-test-retry</id>
+            <activation>
+                <file>
+                    <!--
+                        Activated per-module: only modules containing a .flaky-tests-marker file
+                        get the two-phase surefire execution that retries @Tag("flaky") tests if they fail.
+                    -->
+                    <exists>.flaky-tests-marker</exists>
+                </file>
+            </activation>
+            <build>
+                <plugins>
+                    <plugin>
+                        <artifactId>maven-surefire-plugin</artifactId>
+                        <executions>
+                            <execution>
+                                <id>default-test</id>
+                                <configuration>
+                                    <excludedGroups>flaky</excludedGroups>
+                                </configuration>
+                            </execution>
+                            <execution>
+                                <id>flaky-tests</id>
+                                <goals>
+                                    <goal>test</goal>
+                                </goals>
+                                <configuration>
+                                    <groups>flaky</groups>
+                                    <rerunFailingTestsCount>3</rerunFailingTestsCount>
+                                </configuration>
+                            </execution>
+                        </executions>
+                    </plugin>
+                </plugins>
+            </build>
+        </profile>
     </profiles>
 
 </project>

reference/grpc/.flaky-tests-marker

@@ -0,0 +1,3 @@
+This marker file activates the "flaky-test-retry" profile defined in the root pom.xml.
+When present, tests annotated with @Tag("flaky") are excluded from the default surefire
+execution and run in a separate execution with up to 3 retries on failure.

reference/jsonrpc/.flaky-tests-marker

@@ -0,0 +1,3 @@
+This marker file activates the "flaky-test-retry" profile defined in the root pom.xml.
+When present, tests annotated with @Tag("flaky") are excluded from the default surefire
+execution and run in a separate execution with up to 3 retries on failure.

reference/rest/.flaky-tests-marker

@@ -0,0 +1,3 @@
+This marker file activates the "flaky-test-retry" profile defined in the root pom.xml.
+When present, tests annotated with @Tag("flaky") are excluded from the default surefire
+execution and run in a separate execution with up to 3 retries on failure.

tests/multiversion/grpc/.flaky-tests-marker

@@ -0,0 +1,3 @@
+This marker file activates the "flaky-test-retry" profile defined in the root pom.xml.
+When present, tests annotated with @Tag("flaky") are excluded from the default surefire
+execution and run in a separate execution with up to 3 retries on failure.

tests/multiversion/jsonrpc/.flaky-tests-marker

@@ -0,0 +1,3 @@
+This marker file activates the "flaky-test-retry" profile defined in the root pom.xml.
+When present, tests annotated with @Tag("flaky") are excluded from the default surefire
+execution and run in a separate execution with up to 3 retries on failure.

tests/multiversion/rest/.flaky-tests-marker

@@ -0,0 +1,3 @@
+This marker file activates the "flaky-test-retry" profile defined in the root pom.xml.
+When present, tests annotated with @Tag("flaky") are excluded from the default surefire
+execution and run in a separate execution with up to 3 retries on failure.

tests/server-common/src/test/java/org/a2aproject/sdk/server/apps/common/AbstractA2AServerTest.java

@@ -88,6 +88,7 @@
 import org.a2aproject.sdk.spec.UnsupportedOperationError;
 import org.a2aproject.sdk.spec.UpdateEvent;
 import org.junit.jupiter.api.Assertions;
+import org.junit.jupiter.api.Tag;
 import org.junit.jupiter.api.Test;
 import org.junit.jupiter.api.Timeout;
 
@@ -782,6 +783,7 @@ public void testSendMessageStreamExistingTaskSuccess() throws Exception {
     }
 
     @Test
+    @Tag("flaky")
     @Timeout(value = 3, unit = TimeUnit.MINUTES)
     public void testSubscribeExistingTaskSuccess() throws Exception {
         saveTaskInTaskStore(MINIMAL_TASK);
@@ -1711,6 +1713,7 @@ public void testDeletePushNotificationConfigSetWithoutConfigId() throws Exceptio
     }
 
     @Test
+    @Tag("flaky")
     @Timeout(value = 1, unit = TimeUnit.MINUTES)
     public void testNonBlockingWithMultipleMessages() throws Exception {
         AtomicReference<String> generatedTaskIdRef = new AtomicReference<>();
@@ -3076,6 +3079,7 @@ public void testAgentToAgentDelegation() throws Exception {
      * </ul>
      */
     @Test
+    @Tag("flaky")
     public void testAgentToAgentLocalHandling() throws Exception {
         // No taskId - server generates one; routing is by message content prefix "a2a-local:"
         Message localMessage = Message.builder()