Der Fall wird zweimal erstellt, indem die API in Spring/Java zweimal aufgerufen wird ⇐ Java

[Anonymous]

Wir haben Code wie diesen, der die API im Rahmen des Abfragemechanismus dreimal wiederholt. Falls es nicht abgeschlossen wird, wird ein Fall erstellt, indem eine andere API aufgerufen wird. Was passiert, ist, dass ich in den Protokollen zwei Threads sehe – den Hauptthread und den Thread-Pool, der neue Fälle erstellt. Daher wird die Tracking-ID aus dem Thread-Pool thehead auf null gesetzt und somit doppelte Threads erstellt.
Hier ist ein relevanter Teil des Codes. LLM schlägt mir vor, eine Zeitüberschreitung unter dem Abbruch-Aufruf hinzuzufügen, aber eine andere API, bei der es keine Zeitüberschreitung gibt, ist damit konfrontiert.
Hauptthread

Code: Select all

if (entity != null && entity.getId() != null) {
Future future = null;
try {
log.info("Waiting for polling result for id : {} trackingId: {}", entity.getId(), trackingId);
OrderRequest finalEntity = entity;
String finalProductOrderId = productOrderId;
Callable task = () -> checkGetStatusAPIForCeaseLine(customerId, finalProductOrderId, routeEnv, sc, trackingId, finalEntity, req);
future = executor.submit(task);
CeaseLineResponse result = future.get(futureTimeoutSecondsForCeaseLine, TimeUnit.SECONDS);
return result;
} catch (TimeoutException te) {
log.warn("Polling did not complete in time for id: {} trackingId: {}", entity.getId(), trackingId);
future.cancel(true);
return handleAmaCaseForCeaseLineTimeout(customerId, entity, productOrderId, req, httpHeaders, trackingId);
} catch (Exception e) {
log.error("Error while waiting for poll result", e);
return CeaseLineResponse.builder()
.httpCode(500)
.message("FAILED")
.orderNo(entity.getBillerOrderNo())
.orderActionId(entity.getBillerOrderActionNo())
.build();
}
}

Executor-Aufruf

Code: Select all

@Transactional
public CeaseLineResponse checkGetStatusAPIForCeaseLine(String customerId, String productOrderId, String routeEnv, String sc, String trackingId, OrderRequest entity, CreateAndSubmitSuspendResumeCeaseOrderRequest req) {

HttpHeaders httpHeaders = new HttpHeaders();
httpHeaders.add(HEADER_TRACKING_ID, trackingId);
httpHeaders.add(HEADER_ROUTE_ENV, routeEnv);
httpHeaders.add(HEADER_SOURCE_SYSTEM_ID, "MATRIX-CEASE-POLL");
httpHeaders.add(HEADER_SOURCE_SERVER_ID, "MATRIX-CEASE-POLL");
httpHeaders.add(HEADER_SERVICE_NAME, "OrderStatus");
httpHeaders.add(HEADER_CACHE_CONTROL, HEADER_CACHE_CONTROL_NO_CACHE);
httpHeaders.add(HEADER_SERVICE_OPERATION_NAME, "MATRIX-CEASE-POLL");
httpHeaders.add(HEADER_TIMESTAMP, String.valueOf(System.currentTimeMillis()));

for (int i = 0; i < maxAttemptForCeaseLine; i++) {
// wait before polling
try {
Thread.sleep(ceaseLineSubmittedPlan[i] * 1000);
} catch (InterruptedException e) {
Thread.currentThread().interrupt();
log.info("Polling interrupted due to timeout, trackingId: {}", trackingId);
return null;
}

try {
long start = System.nanoTime();
OrderStatusResponse orderStatus = amdocs.getOrderStatus(customerId, productOrderId, sc, trackingId, httpHeaders);
long end = System.nanoTime();
double durationSeconds = (end - start) / 1_000_000_000.0;
log.info("Polling attempt {} for getCeaseLineOrderStatus.  Execution time: {} seconds, trackingId: {}", i + 1, durationSeconds, trackingId);
String status = orderStatus != null && orderStatus.getMessage() != null ? orderStatus.getMessage().getStatus() : null;
log.info("Polled status: {} for cease line, trackingId: {}", status, trackingId);

if ("CLOSED".equalsIgnoreCase(status)) {
entity.setStatus(OrderStatus.SUCCESS);
entity.setStatusUpdateDate(Instant.now());
entity.setStatusUpdateSource("API");
entity.setAttemptCount(i + 1);
final OrderRequest save = ceaseOrderRepo.save(entity);
log.info("Obtained CLOSED status for cease line. Updated row: {} trackingId: {}", save, entity.getTrackingId());

return CeaseLineResponse.builder()
.httpCode(200)
.message("SUCCESS")
.orderNo(entity.getBillerOrderNo())
.orderActionId(entity.getBillerOrderActionNo())
.build();
}
} catch (Exception ex) {
entity.setStatus(OrderStatus.FAILED);
entity.setRejectReason("Poll error: " + checkError(ex));
entity.setAttemptCount(i + 1);
final OrderRequest save = ceaseOrderRepo.save(entity);
log.info("Obtained FAILED status for cease line. Updated row: {} trackingId: {}", save, entity.getTrackingId());

return CeaseLineResponse.builder()
.httpCode(500)
.message("FAILED")
.orderNo(entity.getBillerOrderNo())
.orderActionId(entity.getBillerOrderActionNo())
.messageDescription(entity.getRejectReason())
.build();
}
}

entity.setAttemptCount(maxAttemptForReplaceDevice);
return handleAmaCaseForCeaseLineTimeout(customerId, entity, productOrderId, req, httpHeaders, trackingId);
}

private CeaseLineResponse handleAmaCaseForCeaseLineTimeout(
String customerId,
OrderRequest entity,
String productOfferId,
CreateAndSubmitSuspendResumeCeaseOrderRequest req,
HttpHeaders httpHeaders,
String trackingId
) {
log.info("Creating case in AMA for stuck cease order. Order ID: {}, Action ID: {}, trackingId: {}", entity.getBillerOrderNo(), entity.getBillerOrderActionNo(),
trackingId);
String bpmOrderId = req.getPayload().getBpmOrderInputX9() != null
? req.getPayload().getBpmOrderInputX9().getBpmOrderIdx9()
: null;
String caseNumber = maybeCreateAmaCase(customerId, entity, productOfferId, bpmOrderId, httpHeaders);

entity.setRejectReason("Order created but still processing - case Number:  " + caseNumber);
entity.setStatus(OrderStatus.SUBMITTED);
final OrderRequest save = ceaseOrderRepo.save(entity);
log.info("SUBMITTED (polling finished) for cease line. Updated row: {} trackingId: {}", save, entity.getTrackingId());

return buildAcceptedOrderResponse202(entity.getBillerOrderNo(), entity.getBillerOrderActionNo(), "Order created but still processing", caseNumber);
}

Dies geschieht in der Anfrage und beim Erstellen eines neuen Falls mithilfe der API

Code: Select all

u_correlation_id\":\"null:9001962483:9001962484

Dies ist, was LLM mir vorgeschlagen hat, nach dem Abbruch-Aufruf hinzuzufügen:

Code: Select all

// Add a small delay to ensure the polling thread has stopped
try {
Thread.sleep(100);
} catch (InterruptedException ie) {
Thread.currentThread().interrupt();
}

und eine unterbrochene Prüfung hier"

Code: Select all

for (int i = 0; i < maxAttemptForCeaseLine; i++) {
// Check if thread has been interrupted/cancelled
if (Thread.currentThread().isInterrupted()) {
log.info("Polling cancelled for trackingId: {}", trackingId);
return null;
}

Was läuft hier schief? Irgendwelche Gedanken?