Fix JMS1 CLIENT_ACKNOWLEDGE flaky test with high scope iteration keep-alive (#10798)

Mark receiving messages with manual acknowledgement as @flaky for JMS1V1ForkedTest

The TEMPORARY_TOPIC variant of this test has a race condition where the 3rd
consumer trace can complete before acknowledge() is called, yielding 6 traces
instead of the expected 5 at the first assertion point. This is inherent to
the async nature of JMS message processing with TEMPORARY_TOPICs.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Fix JMS1 CLIENT_ACKNOWLEDGE flaky test by tolerating early scope cleanup

The "receiving messages from TEMPORARY_TOPIC with manual acknowledgement"
test in JMS1V1ForkedTest was flaky because the scope iteration keep-alive
cleanup (set to 1 second in tests) could finish the 3rd consumer span
before acknowledge() was called, producing 6 traces instead of expected 5.

Root cause: In the legacy context manager path, activateNext() schedules a
root iteration scope cleanup. If this fires before the test's intermediate
assertTraces(5) check, the 3rd consumer trace is already complete.

Fix: Use assertTraces(5, true) for the intermediate assertion to tolerate
additional traces. Also fix ListWriterAssert.ignoreAdditionalTraces to
properly use >= instead of == for the size check, matching the behavior
of the Java TraceAssertions class.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Fix JMS1 CLIENT_ACKNOWLEDGE flaky test with low scope iteration keep-alive

Set SCOPE_ITERATION_KEEP_ALIVE to 1ms in the manual acknowledgement test
so the 3rd consumer span is always cleaned up before assertion, making
the test deterministic with assertTraces(6) instead of tolerating 5 or 6.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Retry CI after infrastructure failures

Retry CI - ghcr.io connectivity issue

ci: retrigger pipeline (infrastructure failures)

fix: restore acknowledge() before assertTraces to fix timeout

The 3rd consumer span needs acknowledge() to complete before
assertTraces(6) can succeed. SCOPE_ITERATION_KEEP_ALIVE alone
is not sufficient — the acknowledge triggers the span to finish.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

fix: restore two-phase assertion with high keep-alive to prevent flakiness

Set SCOPE_ITERATION_KEEP_ALIVE to 10s to prevent early cleanup of the
3rd consumer span. This makes the test deterministic:
- Phase 1: 5 traces (3rd consumer span still open, not yet acknowledged)
- Phase 2: 6 traces (after receivedMessage3.acknowledge())

The original flakiness was caused by the default keep-alive being short
enough that the 3rd consumer span sometimes got cleaned up before the
first assertion.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

ci: retrigger pipeline to resolve SSI infrastructure failures

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Co-authored-by: devflow.devflow-routing-intake <devflow.devflow-routing-intake@kubernetes.us1.ddbuild.io>

Author: bm1549

dd-java-agent/instrumentation/jms/javax-jms-1.1/src/test/groovy/JMS1Test.groovy

@@ -7,6 +7,7 @@ import datadog.trace.agent.test.naming.VersionedNamingTestBase
 import datadog.trace.api.Config
 import datadog.trace.api.DDSpanTypes
 import datadog.trace.api.Trace
+import datadog.trace.api.config.TracerConfig
 import datadog.trace.api.config.TraceInstrumentationConfig
 import datadog.trace.bootstrap.instrumentation.api.InstrumentationTags
 import datadog.trace.bootstrap.instrumentation.api.Tags
@@ -257,6 +258,9 @@ abstract class JMS1Test extends VersionedNamingTestBase {
 
   def "receiving messages from #destinationType with manual acknowledgement"() {
     setup:
+    // Use a long scope iteration keep-alive to prevent early cleanup of the 3rd
+    // consumer span, ensuring exactly 5 traces before acknowledge (not 6).
+    injectSysConfig(TracerConfig.SCOPE_ITERATION_KEEP_ALIVE, "10000")
     def destination = destinationType.create(session)
     def clientSession = connection.createSession(false, Session.CLIENT_ACKNOWLEDGE)
     def producer = session.createProducer(destination)
@@ -276,7 +280,8 @@ abstract class JMS1Test extends VersionedNamingTestBase {
     receivedMessage1.text == messageText1
     receivedMessage2.text == messageText2
     receivedMessage3.text == messageText3
-    // only two consume traces will be finished at this point
+    // only two consume traces will be finished at this point because message 3
+    // has not been acknowledged and the long keep-alive prevents early cleanup
     assertTraces(5) {
       producerTraceWithNaming(it, destination)
       producerTraceWithNaming(it, destination)
@@ -300,7 +305,6 @@ abstract class JMS1Test extends VersionedNamingTestBase {
     }
 
     cleanup:
-    receivedMessage3.acknowledge()
     producer.close()
     consumer.close()
     clientSession.close()