zalando/patroni
Patroni
Patroni is a high availability manager and cluster orchestrator for PostgreSQL. It functions as a failover controller that monitors database health and manages leader transitions to ensure continuous availability and consistency across multiple nodes.
The system relies on a distributed configuration store client to synchronize cluster state and perform leader election. It supports integration with external consensus services such as Etcd, Consul, and ZooKeeper, and provides specific orchestration capabilities for deploying clusters within Kubernetes environments.
The project covers the management of database replication, allowing for the configuration of synchronous and asynchronous streaming. It also includes mechanisms for monitoring replication lag to prevent data loss during failover events.
Features
- Database Cluster Orchestration - Orchestrates high-availability database clusters with automated failover and replication using distributed configuration stores.
- Streaming Configurations - Provides detailed configuration for synchronous and asynchronous streaming and replication lag thresholds.
- High Availability Configurations - Coordinates automatic failover and replication to ensure continuous database availability and data redundancy.
- High Availability Clustering - Ensures continuous database availability through automatic failover and coordinated replication across server nodes.
- Database Failover Management - Monitors database health and automatically triggers leader transitions to prevent service downtime.
- Lag Management - Manages replication lag and streaming configurations to prevent data loss during failover events.
- Distributed Leader Election - Implements coordination mechanisms for leader election using external distributed consensus stores.
- Kubernetes Orchestration - Facilitates the deployment and management of highly available database clusters within Kubernetes environments.
- Distributed Cluster Provisioners - Synchronizes cluster settings across nodes using a distributed configuration store for real-time state management.
- Cluster State Coordinators - Uses external consensus services like Etcd and ZooKeeper to maintain consistent global cluster state and leader election.
- Database Servers - Template for PostgreSQL HA with ZooKeeper or etcd.
- High Availability - Template for managing high availability with consensus stores.
- High Availability and Scaling - Template for PostgreSQL high availability using distributed configuration stores.
