feat: Stale 데이터 비정상 궤적 전환 — 과거 timestamp 수신 시 정보 보존

- CacheBasedVesselTrackDataReader: stale 그룹 스킵 제거 → 전체 통과 + 감지 로깅
- VesselTrackStepConfig: Processor에서 stale time_bucket 감지 시 비정상 궤적으로 전환
  - abnormal_type: stale_timestamp (신규 분류)
  - time_bucket: 현재 5분 버킷으로 오버라이드 (파티션 존재 보장)
  - details: 원본 time_bucket, 지연 시간(분/시), 속도/거리 정보

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

src/main/java/gc/mda/signal_batch/batch/job/VesselTrackStepConfig.java

@@ -35,8 +35,11 @@ import org.springframework.context.annotation.Profile;
 
 import javax.sql.DataSource;
 import java.sql.Timestamp;
+import java.time.Duration;
+import java.time.LocalDateTime;
 import java.util.ArrayList;
 import java.util.Arrays;
+import java.util.HashMap;
 import java.util.List;
 import java.util.Map;
 import java.util.concurrent.ConcurrentHashMap;
@@ -93,9 +96,6 @@ public class VesselTrackStepConfig {
     @Value("${vessel.batch.chunk-size:1000}")
     private int chunkSize;
 
-    @Value("${partition.retention.tables.t_vessel_tracks_5min.retention-days:7}")
-    private int trackRetentionDays;
-    
     @PostConstruct
     public void init() {
         // 5분 Job의 이름을 명시적으로 설정
@@ -116,7 +116,7 @@ public class VesselTrackStepConfig {
     @Bean
     @StepScope
     public CacheBasedVesselTrackDataReader trackDataReader() {
-        return new CacheBasedVesselTrackDataReader(aisTargetCacheManager, trackRetentionDays);
+        return new CacheBasedVesselTrackDataReader(aisTargetCacheManager);
     }
 
     @Bean
@@ -140,8 +140,15 @@ public class VesselTrackStepConfig {
 
             // 3. 강화된 비정상 궤적 필터링 (버킷 내 + 버킷 간 점프 검출)
             List<VesselTrack> filteredTracks = new ArrayList<>();
+            LocalDateTime staleCutoff = LocalDateTime.now().toLocalDate().atStartOfDay();
 
             for (VesselTrack track : tracks) {
+                // Stale 데이터 감지 → 비정상 궤적으로 전환 (정상 집계에서 누락)
+                if (track.getTimeBucket() != null && track.getTimeBucket().isBefore(staleCutoff)) {
+                    saveStaleAbnormalTrack(track);
+                    continue;
+                }
+
                 boolean isAbnormal = false;
                 String abnormalReason = "";
 
@@ -285,6 +292,70 @@ public class VesselTrackStepConfig {
         }
     }
 
+    /**
+     * Stale 데이터(오늘 이전 time_bucket)를 비정상 궤적으로 전환 저장
+     * - time_bucket: 현재 5분 버킷으로 오버라이드 (파티션 존재 보장)
+     * - abnormal_type: stale_timestamp
+     * - details: 원본 time_bucket, 지연 시간(분/시), 속도/거리
+     */
+    private void saveStaleAbnormalTrack(VesselTrack track) {
+        LocalDateTime now = LocalDateTime.now();
+        LocalDateTime currentBucket = now.withSecond(0).withNano(0)
+                .minusMinutes(now.getMinute() % 5);
+        LocalDateTime originalTimeBucket = track.getTimeBucket();
+        long delayMinutes = Duration.between(originalTimeBucket, now).toMinutes();
+
+        VesselTrack staleTrack = VesselTrack.builder()
+                .mmsi(track.getMmsi())
+                .timeBucket(currentBucket)
+                .trackGeom(track.getTrackGeom())
+                .distanceNm(track.getDistanceNm())
+                .avgSpeed(track.getAvgSpeed())
+                .maxSpeed(track.getMaxSpeed())
+                .pointCount(track.getPointCount())
+                .startPosition(track.getStartPosition())
+                .endPosition(track.getEndPosition())
+                .build();
+
+        log.info("Stale 데이터 비정상 전환: MMSI={}, 원본={}, 현재={}, 지연={}분",
+                track.getMmsi(), originalTimeBucket, currentBucket, delayMinutes);
+
+        Map<String, Object> details = new HashMap<>();
+        details.put("originalTimeBucket", originalTimeBucket.toString());
+        details.put("currentTimeBucket", currentBucket.toString());
+        details.put("delayMinutes", delayMinutes);
+        details.put("delayHours", delayMinutes / 60);
+        if (track.getAvgSpeed() != null) details.put("avgSpeed", track.getAvgSpeed());
+        if (track.getDistanceNm() != null) details.put("distanceNm", track.getDistanceNm());
+        if (track.getPointCount() != null) details.put("pointCount", track.getPointCount());
+
+        List<AbnormalTrackDetector.AbnormalSegment> segments = List.of(
+                AbnormalTrackDetector.AbnormalSegment.builder()
+                        .type("stale_timestamp")
+                        .startIndex(0)
+                        .endIndex(track.getPointCount() != null
+                                ? Math.max(track.getPointCount() - 1, 0) : 0)
+                        .actualValue(delayMinutes)
+                        .threshold(0)
+                        .description(String.format("Stale 데이터: 원본 %s, 지연 %d분 (%d시간)",
+                                originalTimeBucket, delayMinutes, delayMinutes / 60))
+                        .details(details)
+                        .build());
+
+        AbnormalDetectionResult result = AbnormalDetectionResult.builder()
+                .originalTrack(staleTrack)
+                .correctedTrack(null)
+                .abnormalSegments(segments)
+                .hasAbnormalities(true)
+                .build();
+
+        try {
+            abnormalTrackWriter.write(new Chunk<>(List.of(result)));
+        } catch (Exception e) {
+            log.error("Stale 비정상 궤적 저장 실패: MMSI={}", track.getMmsi(), e);
+        }
+    }
+
     // CompositeItemWriter로 3개 테이블에 동시 저장
     @Bean
     @StepScope

src/main/java/gc/mda/signal_batch/batch/reader/CacheBasedVesselTrackDataReader.java

@@ -29,7 +29,6 @@ import java.util.stream.Collectors;
 public class CacheBasedVesselTrackDataReader implements ItemReader<List<VesselData>> {
 
     private final AisTargetCacheManager cacheManager;
-    private final int staleDataThresholdDays;
 
     private Iterator<List<VesselData>> groupIterator;
     private boolean initialized = false;
@@ -59,11 +58,12 @@ public class CacheBasedVesselTrackDataReader implements ItemReader<List<VesselDa
         }
 
         // AisTargetEntity → VesselData 변환 + MMSI × 5분 time_bucket 이중 그룹화
-        LocalDateTime staleCutoff = LocalDateTime.now().minusDays(staleDataThresholdDays);
+        // 오늘 00:00 이전 데이터 필터 (파티션은 오늘부터만 존재)
+        LocalDateTime staleCutoff = LocalDateTime.now().toLocalDate().atStartOfDay();
         List<List<VesselData>> allGroups = new ArrayList<>();
         long totalPoints = 0;
         int totalVessels = 0;
-        int skippedStaleGroups = 0;
+        int staleGroups = 0;
 
         for (Map.Entry<String, List<AisTargetEntity>> entry : trackBuffer.entrySet()) {
             List<VesselData> vesselDataList = entry.getValue().stream()
@@ -88,16 +88,15 @@ public class CacheBasedVesselTrackDataReader implements ItemReader<List<VesselDa
 
             for (Map.Entry<LocalDateTime, List<VesselData>> bucketEntry : bucketGroups.entrySet()) {
                 if (bucketEntry.getKey().isBefore(staleCutoff)) {
-                    skippedStaleGroups++;
+                    staleGroups++;
-                    continue;
                 }
                 allGroups.add(bucketEntry.getValue());
                 totalPoints += bucketEntry.getValue().size();
             }
         }
 
-        if (skippedStaleGroups > 0) {
+        if (staleGroups > 0) {
-            log.info("Stale data 필터: {}개 그룹 스킵 ({}일 이전 데이터)", skippedStaleGroups, staleDataThresholdDays);
+            log.info("Stale 그룹 감지: {}건 (기준: {} 이전, 비정상 궤적으로 전환 예정)", staleGroups, staleCutoff);
         }
         log.info("트랙 버퍼 Reader 초기화: {} 선박, {} 그룹(MMSI×버킷), {} 포인트 (평균 {}pt/그룹)",
                 totalVessels, allGroups.size(), totalPoints,