Monitoring

Overview

7 built-in monitoring views and Prometheus metrics export.

Monitoring Views

NeuronDB provides 7 comprehensive monitoring views for real-time observability:

1. vector_stats

Aggregate statistics for all vector operations including total vectors, dimensions, memory usage, and operation counts.

2. index_health

Health dashboard for all indexes with status icons, usage statistics, fragmentation levels, and maintenance recommendations.

3. tenant_quota_usage

Quota monitoring with warnings for multi-tenant deployments. Tracks resource usage per tenant and alerts when approaching limits.

4. llm_job_status

Job queue summary for LLM operations including pending, running, completed, and failed jobs. Tracks job processing times and success rates.

5. query_performance

Performance metrics for the last 24 hours including query latencies, throughput, cache hit rates, and slow query identification.

6. index_maintenance_status

Index maintenance operations tracking including defragmentation progress, rebuild schedules, and maintenance history.

7. metrics_summary

Prometheus-compatible metrics summary with key performance indicators ready for external monitoring systems.

All 7 monitoring views

-- Vector statistics - Aggregate statistics for all vector operations
SELECT * FROM neurondb.vector_stats;

-- Index health dashboard - Health status with icons and recommendations
SELECT * FROM neurondb.index_health;

-- Tenant quota usage - Multi-tenant resource monitoring with warnings
SELECT * FROM neurondb.tenant_quota_usage;

-- LLM job queue status - Job processing summary and statistics
SELECT * FROM neurondb.llm_job_status;

-- Query performance metrics - Last 24 hours performance data
SELECT * FROM neurondb.query_performance;

-- Index maintenance operations - Maintenance tracking and history
SELECT * FROM neurondb.index_maintenance_status;

-- Prometheus metrics summary - Metrics ready for external monitoring
SELECT * FROM neurondb.metrics_summary;

Extension Statistics

Extension statistics

-- Extension statistics
SELECT * FROM pg_stat_neurondb();

-- Worker status
SELECT * FROM neurondb_worker_status();

Prometheus Export

Prometheus metrics are available for external monitoring.

Learn More

For detailed documentation on monitoring, metrics interpretation, and alerting, visit: Monitoring Documentation