High Availability

Running Conexor in production with redundancy

Architecture

For production deployments, Conexor supports horizontal scaling and redundancy:

                    ┌─────────────┐
                    │   LB / CDN  │
                    └──────┬──────┘
                           │
         ┌─────────────────┼─────────────────┐
         │                 │                 │
    ┌────▼────┐      ┌────▼────┐      ┌────▼────┐
    │ Backend │      │ Backend │      │ Backend │
    │   #1    │      │   #2    │      │   #3    │
    └────┬────┘      └────┬────┘      └────┬────┘
         │                │                │
         └────────────────┼────────────────┘
                          │
              ┌───────────┴───────────┐
              │                       │
         ┌────▼────┐            ┌────▼────┐
         │PostgreSQL│            │  Redis  │
         │ Primary  │            │ Cluster │
         └────┬────┘            └─────────┘
              │
         ┌────▼────┐
         │PostgreSQL│
         │  Replica │
         └─────────┘

Backend Scaling

The backend is stateless and can be scaled horizontally:

• Deploy multiple backend instances behind a load balancer
• Sessions are stored in Redis (shared between instances)
• SignalR uses Redis backplane for real-time sync
• No sticky sessions required

Database HA

For PostgreSQL high availability:

• Use streaming replication with a primary + replica
• Configure automatic failover with Patroni or repmgr
• Point backend to a connection pooler (PgBouncer)
• Regular backups with pgBackRest or WAL-G

Redis HA

For Redis high availability:

• Use Redis Sentinel for automatic failover
• Or use Redis Cluster for sharding + HA
• Enable AOF persistence for durability

Agent Redundancy

For critical databases, run multiple agents:

• Install agents on different servers
• Each agent registers independently
• Both can serve queries for the same data source
• Load balancing handled automatically by the cloud service

Monitoring

• Use the /health endpoint for load balancer health checks
• Monitor PostgreSQL replication lag
• Set up alerts for Redis memory usage
• Track agent connection status in the dashboard