Kubernetes Operator¶
The HyperbyteDB Operator is a Kubernetes operator that automates deploying and managing HyperbyteDB database clusters. It extends the Kubernetes API with three custom resources:
| Custom Resource | Purpose |
|---|---|
| HyperbytedbCluster | Declares a HyperbyteDB cluster (single-node or multi-replica) with storage, networking, TLS, monitoring, autoscaling, and failover configuration |
| HyperbytedbBackup | Defines one-shot or scheduled S3 backups of a cluster with configurable retention |
| HyperbytedbRestore | Restores a cluster from a HyperbytedbBackup snapshot |
How It Works¶
The operator watches for these custom resources and reconciles the underlying Kubernetes objects (StatefulSets, Services, PVCs, ConfigMaps, Secrets) to match the desired state. Key capabilities include:
- Rolling upgrades -- Change the
versionfield and the operator upgrades pods one at a time, waiting for each to become healthy before proceeding. - Automatic TLS -- Enable
server.tls.enabledand the operator generates a self-signed CA and per-node certificates, or integrates with cert-manager. - Cluster topology -- For multi-replica clusters, the operator configures Raft consensus, peer discovery, heartbeat, and anti-entropy settings automatically from the StatefulSet ordinals.
- Monitoring integration -- Optionally creates ServiceMonitor resources for Prometheus and Grafana dashboard ConfigMaps.
- Autoscaling -- Configures HorizontalPodAutoscaler based on CPU utilization targets.
- Failover -- Detects unhealthy members and triggers automatic recovery within configurable thresholds.
- Pause/resume -- Set
spec.paused: trueto freeze reconciliation for manual maintenance windows.
Cluster Lifecycle Phases¶
The operator tracks cluster state through these phases:
| Phase | Description |
|---|---|
Pending | Resource created, reconciliation not yet started |
Initializing | StatefulSet and services being created |
Running | All replicas healthy and serving traffic |
Scaling | Replica count change in progress |
Upgrading | Rolling version upgrade in progress |
Failed | Reconciliation encountered an unrecoverable error |
Next Steps¶
- Installation -- Install the operator via Helm
- HyperbytedbCluster -- Full CRD reference and example configurations
- Backup & Restore -- Set up automated backups and perform restores
- hyperbytedb-proxy -- Optional health-aware HTTP reverse proxy in front of the database Service