pgbouncer/pgbouncer
Pgbouncer
PgBouncer is a lightweight middleware proxy server and connection pooler for PostgreSQL. It acts as a database connection manager and load balancer, reducing system overhead by managing a pool of reusable connections and distributing incoming traffic across multiple backend servers.
The project distinguishes itself through a protocol-level proxying approach that handles authentication and TLS encryption without parsing underlying SQL queries. It provides specialized support for mapping named prepared statements across different backend connections and implements a socket-sharing process model to distribute network traffic across multiple CPU cores.
The software covers a broad range of resource management capabilities, including concurrent connection limiting, idle connection timeouts, and various connection assignment strategies. It integrates centralized authentication via LDAP and PAM, offers real-time connection state inspection, and supports zero-downtime process restarts and graceful shutdowns.
The system integrates with systemd on Linux and as a service on Windows for automated lifecycle management.
Features
- PostgreSQL Pool Managers - Provides a specialized connection pool manager tailored for the PostgreSQL wire protocol and server behavior.
- Connection Pooling - Manages a pool of reusable backend connections assigned to clients based on session or transaction boundaries.
- Resource Limits - Prevents database crashes by capping concurrent connections and terminating idle sessions to save system resources.
- Database Connection Managers - Manages the lifecycle of database connections and implements session or transaction-level pooling.
- Connection Pool Managers - Controls minimum and maximum pool sizes to maintain stability during traffic spikes.
- Connection Limit Managers - Caps the maximum number of global or per-user connections to protect system resource capacity.
- PostgreSQL Proxy Servers - Functions as a network intermediary handling authentication, TLS encryption, and routing between clients and PostgreSQL databases.
- Cross-Connection Mapping - Tracks client-side prepared statements and mirrors them across various backend connections for transparent query execution.
- Load Balancers - Balances incoming database traffic across multiple backend hosts using strategies like round-robin to optimize throughput.
- Connection Pooling Strategies - Assigns server connections based on session, transaction, or statement strategies to optimize performance.
- Protocol-Level Proxying - Intercepts and forwards PostgreSQL wire protocol messages between clients and servers without parsing the underlying SQL queries.
- Non-Blocking Event Loops - Manages all network I/O and timers using a non-blocking event loop to handle thousands of concurrent connections efficiently.
- Connection Pool Observability - Provides real-time inspection of active clients and connection pools to debug connectivity and resource usage.
- Idle Connection Cleanup - Automatically closes inactive database connections and terminates long-running transactions to reclaim system resources.
- PostgreSQL Load Balancers - Distributes incoming database traffic across multiple backend PostgreSQL servers to optimize throughput and availability.
- Statement Synchronization - Tracks and rewrites named prepared statements to ensure they remain available across different assigned server connections.
- Dynamic Configuration - Modifies operational parameters and reloads configuration files at runtime without requiring a restart.
- Connection State Resets - Cleanses session variables and temporary state on a backend connection before returning it to the pool for reuse.
- Dynamic Configuration Reloading - Updates internal state and operational parameters at runtime without disconnecting clients or restarting the process.
- Connection Monitors - Tracks whether client connections are active or idle to analyze real-time resource utilization.
- Zero-Downtime Binaries Upgrades - Upgrades the process while maintaining active client connections to prevent application downtime.
- Administrative Access Control - Provides the ability to pause or enable specific databases to manage client access during maintenance windows.
- Multi-Backend Routing - Distributes client requests across multiple database servers residing on different physical hosts for improved scalability.
- Port Sharing Mechanisms - Binds multiple process instances to the same network port to distribute traffic across CPU cores.
- Socket Sharers - Implements a socket-sharing model that binds multiple process instances to one port to distribute traffic across CPU cores.
- Database PAM Authentication - Delegates user authentication to system-level pluggable authentication modules to verify identities using OS credentials.
- Centralized Identity Management - Integrates PostgreSQL access with external identity providers like LDAP and PAM for unified user management.
- User Identity Verification - Validates user credentials using multiple methods including certificates, external files, and directory services.
- LDAP Authentication - Validates user credentials against a centralized LDAP directory server for identity management.
- Transport Layer Encryption - Secures data in transit between the client and the server using Transport Layer Security.
- Network Statistics Tracking - Tracks transaction counts, query volume, and network traffic to analyze system load and bottlenecks.
