e6data Kubernetes Operator¶

The e6data Kubernetes Operator automates the lifecycle of e6data analytics platform components inside a Kubernetes cluster. It turns declarative resources (CRDs) into a running, production-grade e6data deployment with minimal manual wiring.

Instead of hand-crafting Deployments, Services, configuration, and IAM wiring, you express your intent as custom resources:

CRD	Purpose
NamespaceConfig	Shared infrastructure settings (cloud, storage, scheduling) for a namespace
MetadataServices	Storage and Schema services for metadata management
QueryService	Query execution cluster (Planner, Queue, Executor)
E6Catalog	External data catalog registration (Hive, Glue, Unity, Iceberg, Delta)
CatalogRefresh	One-time catalog metadata refresh
CatalogRefreshSchedule	Scheduled catalog refresh (cron-based)
Pool	Shared compute pool for burst capacity
Governance	Data access control policies
GreptimeDBCluster	Time-series database for query history
MonitoringServices	Logs and metrics collection (Vector-based)

The operator continuously reconciles these resources towards the desired state, creating and updating the underlying Kubernetes primitives and cloud integrations.

Architecture Overview¶

┌─────────────────────────────────────────────────────────────────────────────────┐
│                           KUBERNETES CLUSTER                                     │
│                                                                                  │
│  ┌─────────────────────────────────────────────────────────────────────────┐    │
│  │                    e6-operator-system namespace                          │    │
│  │  ┌─────────────────────────────────────────────────────────────────┐    │    │
│  │  │           E6 OPERATOR (Deployment)                               │    │    │
│  │  │  • Watches CRDs across all namespaces                            │    │    │
│  │  │  • Reconciles desired state → actual state                       │    │    │
│  │  │  • Exposes autoscaling API (:8082)                               │    │    │
│  │  │  • Validation webhooks (:9443)                                   │    │    │
│  │  └─────────────────────────────────────────────────────────────────┘    │    │
│  └─────────────────────────────────────────────────────────────────────────┘    │
│                                        │                                         │
│                    Creates & manages   │                                         │
│                    ┌───────────────────┼───────────────────┐                     │
│                    ▼                   ▼                   ▼                     │
│  ┌─────────────────────────────────────────────────────────────────────────┐    │
│  │                    workspace-prod namespace                              │    │
│  │                                                                          │    │
│  │  ┌──────────────────────┐    ┌──────────────────────────────────────┐   │    │
│  │  │   MetadataServices   │    │           QueryService                │   │    │
│  │  │  ┌────────────────┐  │    │  ┌─────────┐  ┌─────────┐            │   │    │
│  │  │  │Storage Service │  │◄───┤  │ Envoy   │  │ Planner │            │   │    │
│  │  │  │(Blue/Green)    │  │    │  │  (xDS)  │──│  + gRPC │            │   │    │
│  │  │  └────────────────┘  │    │  └─────────┘  └────┬────┘            │   │    │
│  │  │  ┌────────────────┐  │    │                    │                 │   │    │
│  │  │  │Schema Service  │  │    │  ┌─────────┐  ┌────┴────┐            │   │    │
│  │  │  │(Blue/Green)    │  │    │  │  Queue  │  │Executors│ ◄── Autoscale  │    │
│  │  │  └────────────────┘  │    │  │Scheduler│  │ (1..N)  │     or Pool│   │    │
│  │  └──────────────────────┘    │  └─────────┘  └─────────┘            │   │    │
│  │                              └──────────────────────────────────────┘   │    │
│  │  ┌──────────────────────┐    ┌──────────────────────────────────────┐   │    │
│  │  │      E6Catalog       │    │           Pool (optional)            │   │    │
│  │  │  • HIVE / GLUE       │    │  • Burst compute capacity            │   │    │
│  │  │  • UNITY / ICEBERG   │    │  • Karpenter NodePool                │   │    │
│  │  │  • DELTA             │    │  • Warmup DaemonSet                  │   │    │
│  │  └──────────────────────┘    └──────────────────────────────────────┘   │    │
│  └─────────────────────────────────────────────────────────────────────────┘    │
│                                        │                                         │
│  ┌─────────────────────────────────────────────────────────────────────────┐    │
│  │                    greptime-system namespace (optional)                  │    │
│  │  ┌──────────────────────────────────────────────────────────────────┐   │    │
│  │  │              GreptimeDBCluster                                    │   │    │
│  │  │  • Frontend (SQL/gRPC/Prometheus endpoints)                       │   │    │
│  │  │  • Datanode (query execution)                                     │   │    │
│  │  │  • Meta (coordination)                                            │   │    │
│  │  │  • ETCD (metadata persistence)                                    │   │    │
│  │  └──────────────────────────────────────────────────────────────────┘   │    │
│  └─────────────────────────────────────────────────────────────────────────┘    │
└─────────────────────────────────────────────────────────────────────────────────┘
                                         │
                          Connects to    │
          ┌──────────────────────────────┼──────────────────────────────┐
          ▼                              ▼                              ▼
┌──────────────────┐        ┌──────────────────┐        ┌──────────────────┐
│  Object Storage  │        │    Metastore     │        │  Query Clients   │
│  ┌────────────┐  │        │  ┌────────────┐  │        │  ┌────────────┐  │
│  │ S3 / GCS   │  │        │  │ AWS Glue   │  │        │  │ JDBC/ODBC  │  │
│  │ Azure Blob │  │        │  │ Hive HMS   │  │        │  │ REST API   │  │
│  │ Linode Obj │  │        │  │ Unity Cat  │  │        │  │ Python SDK │  │
│  └────────────┘  │        │  │ Iceberg    │  │        │  └────────────┘  │
└──────────────────┘        └──────────────────┘        └──────────────────┘

Data Flow¶

User applies CRs → Operator watches and reconciles
MetadataServices creates Storage + Schema services (blue-green deployment)
E6Catalog registers external catalogs with Storage service via HTTP API
QueryService creates query execution stack (Planner → Queue → Executors), traffic routed via Envoy
Executors read data from object storage, query metastore for schema
Pool (optional) provides burst compute via Karpenter-provisioned nodes
GreptimeDB (optional) stores query history and metrics

Quickstart¶

Prerequisites¶

Component	Version	Purpose
Kubernetes	1.24+	Container orchestration
Helm	3.8+	Package management
cert-manager	1.10+	Webhook TLS certificates
Object Storage	-	S3/GCS/Azure Blob for data

For detailed prerequisites and cloud-specific setup, see the Prerequisites Guide and Installation Guide.

For AWS: Follow the complete AWS Onboarding Guide for step-by-step deployment instructions.

Installation¶

# 1. Install cert-manager (required)
kubectl apply -f https://github.com/cert-manager/cert-manager/releases/download/v1.16.2/cert-manager.yaml

# 2. Wait for cert-manager
kubectl wait --for=condition=Available --timeout=120s -n cert-manager \
  deployment/cert-manager deployment/cert-manager-webhook deployment/cert-manager-cainjector

# 3. Install CRDs
kubectl apply -f config/crd/bases/

# 4. Install the operator
helm install e6-operator ./helm/e6-operator \
  --namespace e6-operator-system \
  --create-namespace \
  --set image.repository=us-docker.pkg.dev/e6data-analytics/e6data/e6-operator \
  --set image.tag=1.0.201 \
  --set imagePullSecrets[0].name=gcr-json-key

# 5. Verify installation
kubectl get pods -n e6-operator-system

For production installations with custom values, see Helm Values Reference.

Deploy Your First Workspace¶

# Create namespace
kubectl create namespace workspace-demo

# Deploy MetadataServices
cat <<EOF | kubectl apply -f -
apiVersion: e6data.io/v1alpha1
kind: MetadataServices
metadata:
  name: demo
  namespace: workspace-demo
spec:
  workspace: demo
  tenant: my-company
  storageBackend: s3a://my-data-lake-bucket

  storage:
    imageTag: "3.0.217"
    resources:
      memory: "8Gi"
      cpu: "4"

  schema:
    imageTag: "3.0.217"
    resources:
      memory: "8Gi"
      cpu: "4"
EOF

# Check status
kubectl get mds demo -n workspace-demo -w

Register a Data Catalog¶

cat <<EOF | kubectl apply -f -
apiVersion: e6data.io/v1alpha1
kind: E6Catalog
metadata:
  name: data-lake
  namespace: workspace-demo
spec:
  catalogType: GLUE
  metadataServicesRef: demo
  isDefault: true
  connectionMetadata:
    catalogConnection:
      glueConnection:
        region: us-east-1
EOF

# Check catalog registration
kubectl get e6cat data-lake -n workspace-demo

Deploy Query Service¶

cat <<EOF | kubectl apply -f -
apiVersion: e6data.io/v1alpha1
kind: QueryService
metadata:
  name: demo-cluster
  namespace: workspace-demo
spec:
  alias: demo
  workspace: demo

  planner:
    image:
      tag: "3.0.217"
    resources:
      memory: "4Gi"
      cpu: "2"

  queue:
    image:
      tag: "3.0.217"
    resources:
      memory: "4Gi"
      cpu: "2"

  executor:
    replicas: 2
    image:
      tag: "3.0.217"
    resources:
      memory: "16Gi"
      cpu: "8"
EOF

# Watch deployment
kubectl get qs demo-cluster -n workspace-demo -w

Key Features¶

Blue-Green Deployments¶

All stateful components (MetadataServices, QueryService) use blue-green deployment strategy: - Zero-downtime updates - Automatic rollback on failure - Release history tracking

Auto-Detection & Configuration¶

The operator auto-detects and configures: - Cloud provider from cluster nodes - Storage endpoints from MetadataServices - Java heap settings from resource limits - Service discovery between components

Multi-Cloud Support¶

Cloud	Object Storage	Metastore	Node Provisioning
AWS	S3	Glue	Karpenter
GCP	GCS	Hive/BigQuery	Karpenter
Azure	Blob	Hive/Unity	Karpenter
Linode	S3-compatible	Hive	LKE Pools

Validation Webhooks¶

All CRDs have comprehensive validation: - Required field checks - Format validation (storage paths, resources) - Cross-field validation (cloud/storage compatibility) - Update safety checks (immutable fields)

Next Steps¶

Task	Link
AWS complete deployment	AWS Onboarding Guide
Infrastructure prerequisites	Prerequisites Guide
Detailed installation	Installation Guide
Helm configuration	Helm Values Reference
Understand all CRDs	CRD Catalog
Production setup	MetadataServices
Query cluster deployment	QueryService
Catalog integration	E6Catalog
Burst scaling	Pool
Logs and metrics	MonitoringServices
Query history	Query History Guide
Autoscaling	Autoscaling Guide
RBAC permissions	RBAC Permissions Guide
Status & diagnostics	Status & Diagnostics Guide
Common issues	Troubleshooting

Support¶

GitHub Issues: https://github.com/e6data/e6-operator/issues
Documentation: https://docs.e6data.io
Slack: #e6data-operator on community Slack