Postgres Operator Karşılaştırma — CloudNativePG, Crunchy, Zalando

"K8s'de Postgres yönetiyorsan operator zorunlu. Manuel StatefulSet + Patroni 2018'in yöntemi. 2026'da operator olmadan 'declarative DB' iddia edemezsin. Soru: hangi operator?"

Bu rehber 3 büyük Postgres operator'ünü karşılaştırır: CloudNativePG, Crunchy PGO, Zalando Postgres Operator. 2026 itibarıyla net karar.

---

⚖️ 3 Büyük Operator

Boyut	CloudNativePG	Crunchy PGO	Zalando
Sahibi	EnterpriseDB (CNCF Sandbox)	Crunchy Data	Zalando
License	Apache 2	Apache 2	MIT
HA backend	Postgres replication (no Patroni)	Patroni	Patroni
DCS	K8s API	K8s API	K8s API
Setup	Çok kolay	Orta	Orta
Backup native	Barman	pgBackRest	WAL-E (eski) / WAL-G
Monitoring	Prometheus exporter built-in	pgmonitor (separate)	exporter + Spilo
Connection pooler	PgBouncer integrated	pgBouncer optional	Connection Manager (custom)
Cluster image	Stock Postgres	Crunchy custom	Spilo (Patroni + Postgres)
Major version upgrade	✅ In-place	✅ Manual + tooling	✅
Multi-cluster (DR)	✅ Native	✅	⚠️ Manual
Active-passive replication	✅	✅	✅
Active-active	❌	❌ (PostgreSQL native değil)	❌
GitOps friendly	✅ Native	✅	✅
Topluluk	Yükselişte	Köklü, enterprise	Köklü, OSS
2026 Öneri	✅✅ K8s ekosisteminde 1.tercih	✅ Enterprise + ticari destek	⚠️ Eski projeler için

---

🌳 Karar Ağacı

START
  │
  ├── Yeni cluster + greenfield?
  │     │
  │     └── EVET → CloudNativePG
  │            (en modern, en kolay, CNCF momentum)
  │
  ├── Enterprise support gerekli + ticari kontrat?
  │     │
  │     └── EVET → Crunchy PGO (Crunchy Data ticari)
  │            veya CloudNativePG (EnterpriseDB ticari)
  │
  ├── Mevcut Patroni setup'ından migrate?
  │     │
  │     └── EVET → Zalando (Patroni native) veya Crunchy
  │
  └── Multi-cluster DR critical?
         │
         └── CloudNativePG (native multi-cluster) veya Crunchy

🎯 2026 net önerisi: CloudNativePG çoğu durum için. Enterprise + premium support ihtiyacı varsa Crunchy.

---

🚀 CloudNativePG — Detay

Niye 2026 önerisi?

• K8s-native: K8s API DCS olarak (Patroni / etcd yok)
• Operator pattern: Tam declarative
• Backup native: Barman + S3
• Monitoring native: Prometheus exporter
• Rolling updates: Zero-downtime
• CNCF Sandbox: Bağımsız topluluk

Install

helm install cnpg cloudnative-pg/cloudnative-pg \
  -n cnpg-system --create-namespace

Cluster manifest (production-ready)

apiVersion: postgresql.cnpg.io/v1
kind: Cluster
metadata:
  name: postgres-prod
  namespace: postgres
spec:
  instances: 3
  imageName: ghcr.io/cloudnative-pg/postgresql:16.4

  postgresql:
    parameters:
      max_connections: "200"
      shared_buffers: "4GB"
      effective_cache_size: "12GB"
      work_mem: "16MB"
      maintenance_work_mem: "1GB"
      synchronous_commit: "on"
      synchronous_standby_names: "ANY 1 (*)"
      log_min_duration_statement: "500ms"
      pg_stat_statements.track: "all"

    pg_hba:
      - host app app 10.0.0.0/8 scram-sha-256
      - host replication replicator 10.0.0.0/8 scram-sha-256

    shared_preload_libraries:
      - pg_stat_statements

  bootstrap:
    initdb:
      database: app
      owner: app
      secret:
        name: postgres-app-creds
      postInitSQL:
        - CREATE EXTENSION IF NOT EXISTS pgcrypto;

  storage:
    size: 100Gi
    storageClass: <FAST_SSD>

  resources:
    requests: {cpu: "2", memory: "8Gi"}
    limits: {cpu: "4", memory: "16Gi"}

  affinity:
    podAntiAffinityType: required
    topologyKey: kubernetes.io/hostname

  monitoring:
    enablePodMonitor: true

  backup:
    barmanObjectStore:
      destinationPath: s3://<BUCKET>/postgres
      s3Credentials:
        accessKeyId: {name: backup-creds, key: ACCESS_KEY}
        secretAccessKey: {name: backup-creds, key: SECRET_KEY}
      wal:
        compression: gzip
        encryption: AES256
      data:
        compression: gzip
        encryption: AES256
    retentionPolicy: "30d"

  certificates:
    serverTLSSecret: postgres-server-cert
    clientCASecret: postgres-client-ca

Otomatik service'ler

postgres-prod-rw  → primary (read-write)
postgres-prod-ro  → replicas (read-only)
postgres-prod-r   → primary + replicas (any)

Failover

# Primary pod'u sil → otomatik failover
kubectl delete pod postgres-prod-1 -n postgres

# 30 saniye içinde yeni primary up
kubectl get cluster postgres-prod -n postgres \
  -o jsonpath='{.status.currentPrimary}'

Backup tetikle

kubectl cnpg backup postgres-prod --backup-name now -n postgres

PITR restore

apiVersion: postgresql.cnpg.io/v1
kind: Cluster
metadata:
  name: postgres-restored
spec:
  bootstrap:
    recovery:
      backup:
        name: backup-20260504
      recoveryTarget:
        targetTime: "2026-05-04 14:30:00.00000+00"

Major version upgrade

spec:
  imageName: ghcr.io/cloudnative-pg/postgresql:17  # 16 → 17

→ Operator rolling upgrade yapar; downtime ~5-10 saniye (failover).

---

🏛️ Crunchy PGO — Enterprise Tier

Niye Crunchy?

• 10+ yıllık enterprise pedigree
• Ticari destek: Crunchy Data subscription
• pgBackRest (en güçlü Postgres backup)
• PostgreSQL Enterprise Manager UI
• TDE, audit, FIPS 140-2 compliance
• Air-gapped deployment

Cluster manifest (örnek)

apiVersion: postgres-operator.crunchydata.com/v1beta1
kind: PostgresCluster
metadata:
  name: hippo
spec:
  postgresVersion: 16
  instances:
    - name: instance1
      replicas: 3
      dataVolumeClaimSpec:
        accessModes: ["ReadWriteOnce"]
        resources:
          requests:
            storage: 100Gi
  backups:
    pgbackrest:
      repos:
        - name: repo1
          s3:
            bucket: <BUCKET>
            endpoint: s3.amazonaws.com
            region: <REGION>

Crunchy K8s'de Crunchy custom container'ları kullanır (stock Postgres image değil).

---

🌍 Zalando Postgres Operator

Niye Zalando?

• Patroni'yi K8s'e adaptasyon
• Spilo image (Patroni + Postgres + WAL-G önceden installed)
• Açık kaynak, stable
• Mevcut Patroni-based migration için doğal

Manifest

apiVersion: acid.zalan.do/v1
kind: postgresql
metadata:
  name: my-postgres
spec:
  teamId: "<TEAM>"
  volume:
    size: 100Gi
  numberOfInstances: 3
  users:
    app:
      - superuser
      - createdb
  databases:
    app: app
  postgresql:
    version: "16"

🔑 Zalando son yıllarda yavaşladı — Crunchy ve CloudNativePG daha aktif.

---

🚧 Operator Olmayan Senaryolar

"DB managed servisi kullan" zaten önerimiz

Bkz Postgres-Production-Guide.md.

Senaryo	Tercih
AWS-native + low ops	RDS / Aurora
GCP	CloudSQL / AlloyDB
Multi-cloud	CloudNativePG K8s
On-prem	CloudNativePG / Crunchy
Air-gapped	Crunchy enterprise

---

🚫 Anti-Pattern Tablosu

Anti-pattern	Niye kötü	Doğru
Manual StatefulSet + Patroni	Operasyon yükü	Operator (CNPG / Crunchy)
Operator kullanıyor ama backup yok	İhlal an meselesi	Barman / pgBackRest auto-config
Single replica "yeter"	SPOF	min 3 replica HA
Sync replication yok	RPO yüksek	synchronous_standby_names="ANY 1 (*)"
Pod anti-affinity yok	Aynı node'da pod'lar → node down felaket	required + topologyKey: hostname
TLS sertifika manuel	Manuel rotation	cert-manager + operator integration
Major version skip	Direct 13 → 17 (test'siz)	13 → 14 → 15 → 16 → 17
Operator restart manuel	Drift	GitOps + ArgoCD self-heal
Backup sadece cluster içi	Region down	Cross-region replication
Operator upgrade ihmal	Eski operator → bug birikir	Quarterly minor upgrade

---

📋 Operator Adoption Checklist

[ ] Operator seçimi: CNPG / Crunchy / Zalando (karar belgelenmiş)
[ ] HA: 3+ instance (sync + async mix)
[ ] Pod anti-affinity required + topologyKey: hostname
[ ] Storage: fast SSD class, 100Gi+ headroom
[ ] Resources: explicit requests/limits
[ ] postgresql parameters tune edilmiş
[ ] pg_stat_statements + pg_hba doğru
[ ] Backup: S3 + encryption + cross-region
[ ] Restore drill quarterly
[ ] PITR test edilmiş
[ ] Monitoring: postgres-exporter + alarmlar
[ ] TLS: cert-manager + secret rotation
[ ] PgBouncer integrate (CNPG'de native)
[ ] Operator upgrade prosedürü dokumante
[ ] Major version upgrade plan
[ ] DR: cross-region cluster (varsa)
[ ] Documentation: operator + Postgres tunables

---

📚 Referanslar

• CloudNativePG — cloudnative-pg.io
• Crunchy PGO — crunchydata.com/products/crunchy-postgres-for-kubernetes
• Zalando Postgres Operator — postgres-operator.readthedocs.io
• Patroni — patroni.readthedocs.io (operator olmadan)
• CNCF Sandbox — cncf.io
• Postgres-Production-Guide.md
• HA-Patroni-Stolon.md
• Backup-Restore-Patterns.md
• Connection-Pooling.md
• Monitoring-Postgres.md
• Zero-Downtime-Migrations.md

---

"Operator olmadan K8s'de Postgres yönetmek 'declarative' iddiası imkansız — her başarısız failover manuel müdahale gerektirir. Operator = saatlerce manuel iş dakikalara, bus factor 1'den N'e."