Published first release of pgbalancer, an AI-powered PostgreSQL load balancer with REST API and MQTT. Pgbalancer provides connection pooling and load balancing for PostgreSQL, is managed over HTTP, and supports real-time event streaming. The project was created as a fork of pgpool-II with major improvements for smart request routing. The code is written in C and distributed under the PostgreSQL open license.
Main capabilities and features:
- Connection pooling and load balancing.
- Automatic failover with watchdog support.
- Health monitoring and caching requests.
- SSL/TLS, PAM/LDAP authentication.
- Compatible with PostgreSQL 13-18.
- Part of the pgElephant utilities that provide high availability. Can integrate with pgraft for Raft consensus, supports pgSentinel monitoring.
- AI-powered load balancing.
- Machine learning algorithms for smart routing of requests.
- Training is adjusted depending on request response time and the state of server resources.
- Evaluates the state in real-time and predicts the balancing of requests.
- Configurable weight parameters for learning speed and guess rate of random events.
- Can be integrated with Prometheus and monitoring platforms Grafana.
Management via REST API.
- 17 HTTP/JSON functions for complete cluster management.
- Integrated API server running on port 8080.
- Authentication with JWT HMAC-SHA256 tokens.
- Get runtime
/Reports, release notes, official announcements.