Operational Architecture
Dnaerys is engineered for resilience in distributed environments. We treat operations as a core requirement for maintaining a high-performance genomic variant store.
Deployment and Orchestration
Kubernetes is the reference platform for production deployments
- Traffic & Security: All services and pods are behind an Ingress Controller. It acts as Policy Enforcement Point, handling the heavy lifting of authentication and authorization before a request reaches the database nodes.
- Example Setup: To see a production configuration, check Helm chart
Observability in Practice
In control theory, Observability is the measure of how well a system's internal state can be inferred from its external outputs. In distributed systems it has the similar meaning, though definitions may vary.
Dnaerys is built with internal instrumentation, exposing the metrics necessary to monitor the SRE "Four Golden Signals" (Latency, Traffic, Errors, and Saturation).
Monitoring & Health
Dnaerys is integrated with modern cloud-native monitoring:
- Dashboards & Alerts: Integration with Prometheus and Grafana for real-time dashboarding and alerting
- Cluster Reliability: Liveness and readiness probes ensure Kubernetes routes traffic to healthy nodes
- Diagnostic Endpoints: Dedicated REST endpoints for health checks and cluster status
- Performance metrics: Every response carries timestamps and performance metrics
Deep Introspection
For troubleshooting complex issues or JVM behavior, we support:
- Granular Logging: Multi-level logging (Error, Warning, Info, Debug)
- JVM Profiling & Tools: JMX & JVMTI
Resilience
The system is designed to degrade gracefully. If a node fails, the cluster continues to serve requests using the remaining data, clearly marking responses as "potentially incomplete" so your downstream applications can handle the partial state with full context.