apiVersion: v1
kind: Pod
metadata:
name: my-pod
spec:
restartPolicy: Always   # ← 바로 이 부분!
containers:
- name: my-container
image: nginx:latest

apiVersion: v1
kind: Pod
metadata:
name: always-restart-pod
spec:
restartPolicy: Always
containers:
- name: web-server
image: nginx:1.25
ports:
- containerPort: 80

apiVersion: v1
kind: Pod
metadata:
name: batch-job-pod
spec:
restartPolicy: OnFailure
containers:
- name: batch-processor
image: my-batch-app:1.0
command: ["python", "process_data.py"]
env:
- name: DATA_SOURCE
value: "s3://my-bucket/data"

apiVersion: v1
kind: Pod
metadata:
name: nightly-batch
spec:
restartPolicy: OnFailure
containers:
- name: data-processor
image: data-processor:latest
command:
- /bin/sh
- -c
- |
python process.py && echo "성공!" && exit 0
# 실패 시 exit 1 반환 → 재시작
# 성공 시 exit 0 반환 → 재시작 안 함

apiVersion: v1
kind: Pod
metadata:
name: db-migration-pod
spec:
restartPolicy: OnFailure
containers:
- name: migrator
image: flyway:9.0
command: ["flyway", "migrate"]
env:
- name: FLYWAY_URL
value: "jdbc:postgresql://db:5432/mydb"
- name: FLYWAY_USER
valueFrom:
secretKeyRef:
name: db-secret
key: username

apiVersion: batch/v1
kind: Job
metadata:
name: data-migration-job
spec:
completions: 1        # 성공적으로 완료해야 할 횟수
parallelism: 1        # 동시에 실행할 Pod 수
backoffLimit: 4       # 최대 재시도 횟수
template:
spec:
restartPolicy: OnFailure  # Job에서는 OnFailure 또는 Never만 가능!
containers:
- name: migrator
image: my-migrator:1.0
command: ["./migrate.sh"]

spec:
backoffLimit: 6  # 기본값은 6
# 6번 실패하면 Job 자체가 Failed 상태가 됨

apiVersion: v1
kind: Pod
metadata:
name: debug-pod
spec:
restartPolicy: Never
containers:
- name: debugger
image: busybox:latest
command: ["sh", "-c", "env && ls -la /app && cat /app/config.yaml"]

# 환경 변수 덤프용 디버그 Pod
apiVersion: v1
kind: Pod
metadata:
name: env-debugger
namespace: production
spec:
restartPolicy: Never
containers:
- name: debugger
image: alpine:latest
command: ["sh", "-c", "printenv | sort > /tmp/env.txt && cat /tmp/env.txt"]
envFrom:
- configMapRef:
name: app-config
- secretRef:
name: app-secrets

apiVersion: batch/v1
kind: Job
metadata:
name: one-time-task
spec:
backoffLimit: 0  # 재시도 없음
template:
spec:
restartPolicy: Never  # Pod 레벨에서도 재시작 없음
containers:
- name: task
image: my-task:1.0
command: ["./run-once.sh"]

# 컨테이너 상태 확인
kubectl describe pod my-pod

# 출력 예시:
# Containers:
#   my-container:
#     State:          Terminated
#       Reason:       Error
#       Exit Code:    1
#       Started:      Mon, 01 Jan 2024 10:00:00 +0000
#       Finished:     Mon, 01 Jan 2024 10:00:05 +0000
#     Last State:     Terminated
#       Reason:       Error
#       Exit Code:    1
#     Ready:          False
#     Restart Count:  3

apiVersion: apps/v1
kind: Deployment
metadata:
name: web-api-deployment
namespace: production
labels:
app: web-api
version: "1.0"
spec:
replicas: 3
selector:
matchLabels:
app: web-api
template:
metadata:
labels:
app: web-api
spec:
restartPolicy: Always  # Deployment는 Always만 가능
containers:
- name: api-server
image: my-api:1.0.0
ports:
- containerPort: 8080
resources:
requests:
memory: "256Mi"
cpu: "250m"
limits:
memory: "512Mi"
cpu: "500m"
livenessProbe:
httpGet:
path: /health
port: 8080
initialDelaySeconds: 30
periodSeconds: 10
failureThreshold: 3
readinessProbe:
httpGet:
path: /ready
port: 8080
initialDelaySeconds: 10
periodSeconds: 5

apiVersion: batch/v1
kind: Job
metadata:
name: monthly-report-job
namespace: batch
spec:
completions: 1
parallelism: 1
backoffLimit: 3          # 최대 3번 재시도
activeDeadlineSeconds: 3600  # 1시간 내에 완료 안 되면 실패
template:
metadata:
labels:
job-name: monthly-report
spec:
restartPolicy: OnFailure
containers:
- name: report-generator
image: report-gen:2.1
command: ["python", "-u", "generate_report.py"]
env:
- name: REPORT_MONTH
value: "2024-01"
- name: DB_HOST
valueFrom:
secretKeyRef:
name: db-credentials
key: host
resources:
requests:
memory: "512Mi"
cpu: "500m"
limits:
memory: "1Gi"
cpu: "1000m"
restartPolicy: OnFailure

apiVersion: batch/v1
kind: CronJob
metadata:
name: daily-cleanup
namespace: maintenance
spec:
schedule: "0 2 * * *"     # 매일 새벽 2시
concurrencyPolicy: Forbid  # 이전 작업이 끝나지 않으면 새 작업 시작 안 함
successfulJobsHistoryLimit: 3
failedJobsHistoryLimit: 1
jobTemplate:
spec:
backoffLimit: 2
template:
spec:
restartPolicy: OnFailure
containers:
- name: cleanup-worker
image: cleanup:latest
command:
- /bin/sh
- -c
- |
echo "정리 작업 시작: $(date)"
python cleanup_old_data.py --days 90
echo "정리 작업 완료: $(date)"

apiVersion: v1
kind: Pod
metadata:
name: debug-session
namespace: production
annotations:
purpose: "임시 디버깅 - 사용 후 삭제 필요"
spec:
restartPolicy: Never
serviceAccountName: debug-sa
containers:
- name: debugger
image: nicolaka/netshoot:latest  # 네트워크 디버깅 도구 포함
command: ["sleep", "3600"]       # 1시간 동안 유지
securityContext:
capabilities:
add: ["NET_ADMIN", "SYS_PTRACE"]
tolerations:
- operator: "Exists"  # 모든 노드에 스케줄 가능
# 사용 후: kubectl delete pod debug-session -n production

apiVersion: v1
kind: Pod
metadata:
name: app-with-init
spec:
restartPolicy: Always
initContainers:
- name: wait-for-db
image: busybox:latest
command:
- sh
- -c
- |
until nc -z db-service 5432; do
echo "DB 대기 중..."
sleep 2
done
echo "DB 연결 확인!"
- name: run-migrations
image: flyway:9.0
command: ["flyway", "migrate"]
env:
- name: FLYWAY_URL
value: "jdbc:postgresql://db-service:5432/mydb"
containers:
- name: app
image: my-app:1.0
ports:
- containerPort: 8080

# 백오프 계산 예시
1번째 재시작: 2^0 × 10 = 10초
2번째 재시작: 2^1 × 10 = 20초
3번째 재시작: 2^2 × 10 = 40초
4번째 재시작: 2^3 × 10 = 80초
5번째 재시작: 2^4 × 10 = 160초
6번째 재시작: 2^5 × 10 = 320초 → 300초(최대)로 제한
7번째 이후: 계속 300초 간격

# 현재 컨테이너 로그
kubectl logs pod-name -c container-name

# 이전 컨테이너 로그 (크래시 직전 로그)
kubectl logs pod-name -c container-name --previous

# 실시간 로그 스트리밍
kubectl logs pod-name -f

# 마지막 100줄만
kubectl logs pod-name --tail=100

kubectl describe pod pod-name

# Events 섹션에서 확인:
# Warning  BackOff    2m    kubelet  Back-off restarting failed container
# Normal   Pulled     5m    kubelet  Successfully pulled image
# Warning  Failed     5m    kubelet  Error: OOMKilled

# 디버깅을 위해 원래 명령 대신 sleep으로 교체
kubectl patch pod crashing-pod -p '{"spec":{"containers":[{"name":"app","command":["sleep","infinity"]}]}}'

# 또는 새 Pod를 sleep으로 생성
kubectl run debug-pod --image=my-app:1.0 --command -- sleep infinity

# 그 다음 exec로 들어가서 직접 확인
kubectl exec -it debug-pod -- /bin/sh

# OOMKilled 확인
kubectl describe pod pod-name | grep -A5 "Last State"

# 출력 예시:
# Last State:     Terminated
#   Reason:       OOMKilled  ← 메모리 부족!
#   Exit Code:    137
#
# 해결: resources.limits.memory 값을 늘려줘

apiVersion: v1
kind: Pod
metadata:
name: robust-app
spec:
restartPolicy: Always
containers:
- name: app
image: my-app:1.0
ports:
- containerPort: 8080

# 시작 체크: 최대 5분(30초 × 10회) 대기
startupProbe:
httpGet:
path: /startup
port: 8080
failureThreshold: 30
periodSeconds: 10

# 생존 체크: 실패 시 컨테이너 재시작
livenessProbe:
httpGet:
path: /health
port: 8080
initialDelaySeconds: 0  # startupProbe 성공 후 시작
periodSeconds: 10
failureThreshold: 3     # 3번 연속 실패 시 재시작
timeoutSeconds: 5

# 준비 체크: 실패 시 서비스에서 제외 (재시작 없음)
readinessProbe:
httpGet:
path: /ready
port: 8080
periodSeconds: 5
failureThreshold: 3

# Liveness Probe 실패 이벤트 확인
kubectl describe pod my-pod

# 이벤트 예시:
# Warning  Unhealthy  5m  kubelet  Liveness probe failed: HTTP probe failed with statuscode: 503
# Normal   Killing    5m  kubelet  Container app failed liveness probe, will be restarted
# Normal   Pulled     4m  kubelet  Container image "my-app:1.0" already present on machine
# Normal   Created    4m  kubelet  Created container app
# Normal   Started    4m  kubelet  Started container app

# ❌ 잘못된 예시 - 작업 완료 후 무한 재시작!
apiVersion: v1
kind: Pod
metadata:
name: data-processor
spec:
# restartPolicy 없음 → 기본값 Always 적용!
containers:
- name: processor
image: my-processor:1.0
command: ["python", "process.py"]

---
# ✅ 올바른 예시
apiVersion: v1
kind: Pod
metadata:
name: data-processor
spec:
restartPolicy: OnFailure  # 또는 Never
containers:
- name: processor
image: my-processor:1.0
command: ["python", "process.py"]

# OOMKilled 확인 방법
kubectl describe pod my-pod | grep -A 10 "Last State"

# 출력:
# Last State:     Terminated
#   Reason:       OOMKilled
#   Exit Code:    137    ← 128 + 9(SIGKILL) = 137
#   Started:      ...
#   Finished:     ...

# 해결: 메모리 제한 증가
spec:
containers:
- name: app
resources:
limits:
memory: "1Gi"  # 기존 512Mi에서 증가
requests:
memory: "512Mi"

# ❌ 위험한 예시 - 이메일이 여러 번 발송될 수 있음!
apiVersion: batch/v1
kind: Job
metadata:
name: send-newsletter
spec:
template:
spec:
restartPolicy: OnFailure  # 실패 시 재시도 → 이메일 중복 발송!
containers:
- name: mailer
image: mailer:1.0

---
# ✅ 안전한 예시 - 멱등성 없는 작업은 Never 사용
apiVersion: batch/v1
kind: Job
metadata:
name: send-newsletter
spec:
backoffLimit: 0  # 재시도 없음
template:
spec:
restartPolicy: Never  # 실패해도 재시작 없음
containers:
- name: mailer
image: mailer:1.0
# 앱 레벨에서 중복 방지 로직 구현 필요

# ❌ 너무 공격적인 설정 - 부하 시 불필요한 재시작 발생
livenessProbe:
httpGet:
path: /health
port: 8080
initialDelaySeconds: 5   # 너무 짧음
periodSeconds: 5          # 너무 자주
failureThreshold: 1       # 1번만 실패해도 재시작!
timeoutSeconds: 1         # 너무 짧음

---
# ✅ 적절한 설정
livenessProbe:
httpGet:
path: /health
port: 8080
initialDelaySeconds: 30   # 앱 시작 시간 고려
periodSeconds: 10          # 10초마다 체크
failureThreshold: 3        # 3번 연속 실패 시 재시작
timeoutSeconds: 5          # 5초 타임아웃

# PDB: 최소 2개의 Pod가 항상 실행되도록 보장
apiVersion: policy/v1
kind: PodDisruptionBudget
metadata:
name: web-pdb
spec:
minAvailable: 2
selector:
matchLabels:
app: web-api
# 이건 restartPolicy와 별개로 동작해!

apiVersion: v1
kind: Pod
metadata:
name: app-with-sidecar
spec:
restartPolicy: Always  # 모든 컨테이너에 적용
containers:
# 메인 앱
- name: main-app
image: my-app:1.0
ports:
- containerPort: 8080

# 로그 수집 사이드카
- name: log-collector
image: fluentd:v1.16
volumeMounts:
- name: app-logs
mountPath: /var/log/app

# 프록시 사이드카 (Envoy)
- name: envoy-proxy
image: envoyproxy/envoy:v1.28
ports:
- containerPort: 9901

volumes:
- name: app-logs
emptyDir: {}

# 1단계: 데이터 추출
apiVersion: batch/v1
kind: Job
metadata:
name: step1-extract
spec:
template:
spec:
restartPolicy: OnFailure
containers:
- name: extractor
image: etl:1.0
command: ["python", "extract.py"]

---
# 2단계: 데이터 변환 (step1 완료 후 실행)
apiVersion: batch/v1
kind: Job
metadata:
name: step2-transform
spec:
template:
spec:
restartPolicy: OnFailure
containers:
- name: transformer
image: etl:1.0
command: ["python", "transform.py"]

---
# 3단계: 데이터 로드 (step2 완료 후 실행)
apiVersion: batch/v1
kind: Job
metadata:
name: step3-load
spec:
template:
spec:
restartPolicy: Never  # 멱등성 없는 작업
containers:
- name: loader
image: etl:1.0
command: ["python", "load.py"]

---
# 1. Deployment: 항상 3개 레플리카 유지
apiVersion: apps/v1
kind: Deployment
metadata:
name: resilient-app
spec:
replicas: 3
strategy:
type: RollingUpdate
rollingUpdate:
maxSurge: 1
maxUnavailable: 0
template:
spec:
restartPolicy: Always
containers:
- name: app
image: my-app:1.0
livenessProbe:
httpGet:
path: /health
port: 8080
periodSeconds: 10
failureThreshold: 3
readinessProbe:
httpGet:
path: /ready
port: 8080
periodSeconds: 5

---
# 2. PodDisruptionBudget: 최소 2개 항상 유지
apiVersion: policy/v1
kind: PodDisruptionBudget
metadata:
name: resilient-app-pdb
spec:
minAvailable: 2
selector:
matchLabels:
app: resilient-app

---
# 3. HorizontalPodAutoscaler: 부하에 따라 자동 스케일
apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
name: resilient-app-hpa
spec:
scaleTargetRef:
apiVersion: apps/v1
kind: Deployment
name: resilient-app
minReplicas: 3
maxReplicas: 10
metrics:
- type: Resource
resource:
name: cpu
target:
type: Utilization
averageUtilization: 70

# Prometheus 쿼리: Pod 재시작 횟수
kube_pod_container_status_restarts_total

# 특정 네임스페이스의 재시작 횟수
kube_pod_container_status_restarts_total{namespace="production"}

# 지난 1시간 동안 재시작 횟수 증가
increase(kube_pod_container_status_restarts_total[1h])

# 재시작 횟수가 5 이상인 Pod 알림
kube_pod_container_status_restarts_total > 5

# Prometheus AlertManager 규칙 예시
groups:
- name: pod-restart-alerts
rules:
- alert: PodCrashLooping
expr: |
increase(kube_pod_container_status_restarts_total[15m]) > 3
for: 5m
labels:
severity: warning
annotations:
summary: "Pod {{ $labels.pod }} 크래시 루프 감지"
description: "{{ $labels.namespace }}/{{ $labels.pod }}의 {{ $labels.container }} 컨테이너가 15분 내 3번 이상 재시작됨"

- alert: PodFrequentlyRestarting
expr: |
kube_pod_container_status_restarts_total > 10
for: 1m
labels:
severity: critical
annotations:
summary: "Pod {{ $labels.pod }} 빈번한 재시작"
description: "총 재시작 횟수: {{ $value }}"

# 모든 Pod의 재시작 횟수 확인
kubectl get pods --all-namespaces | sort -k5 -rn

# 재시작 횟수가 많은 Pod 찾기
kubectl get pods -A -o custom-columns=\
'NAMESPACE:.metadata.namespace,NAME:.metadata.name,RESTARTS:.status.containerStatuses[0].restartCount' \
| sort -k3 -rn | head -20

# 특정 Pod의 상세 재시작 정보
kubectl get pod my-pod -o jsonpath='{.status.containerStatuses[*].restartCount}'

# 재시작 중인 Pod만 필터링
kubectl get pods -A --field-selector=status.phase=Running | grep -v "0/\|1/1\|2/2\|3/3"