Metrics & Monitoring

Monitoring Options

REST API Metrics

pgBalancer REST API provides real-time metrics on port 8080 (configurable):

curl -s http://localhost:8080/api/v1/backends | jq

curl -s http://localhost:8080/api/v1/pool/summary | jq

bctl CLI Monitoring

Use bctl command-line tool for real-time monitoring and statistics:

# View all backend nodes
bctl node-status

# Table format (human-readable)
bctl node-status --format=table

# JSON format (for scripts)
bctl node-status --format=json

# View connection pool status
bctl pool-status --format=table

# Get comprehensive statistics
bctl stats --format=table

# Process count
bctl proc-count

Prometheus Integration

Export pgBalancer metrics to Prometheus using postgres_exporter or custom scraping:

# prometheus.yml
scrape_configs:
  # pgBalancer REST API metrics
  - job_name: 'pgbalancer'
    scrape_interval: 15s
    static_configs:
      - targets: ['localhost:8080']
    metrics_path: '/api/stats'

Grafana Dashboards

Create Grafana dashboards to visualize pgBalancer metrics:

MQTT Event Monitoring

Monitor pgBalancer events in real-time using MQTT:

# Subscribe to all pgBalancer events
mosquitto_sub -h localhost -t 'pgbalancer/#' -v

# Subscribe to specific event types
mosquitto_sub -h localhost -t 'pgbalancer/node/status' -v
mosquitto_sub -h localhost -t 'pgbalancer/failover' -v

Logging and Log Analysis

# View pgBalancer logs
tail -f /var/log/pgbalancer/pgbalancer.log

# Filter for errors only
tail -f /var/log/pgbalancer/pgbalancer.log | grep ERROR

# View systemd logs
sudo journalctl -u pgbalancer -f

Performance Monitoring Queries

# Monitor AI health scores in real-time
watch -n 5 'curl -s http://localhost:8080/api/backends | jq ".[] | {id, hostname, health_score, avg_response_time_ms}"'

Alerting Best Practices

Key Metrics Reference