Production Deployment & Mesh Replication

What you'll learn

Deploy SapixDB to production with proper configuration, enable mesh replication for high availability, and apply the full production checklist.

Production docker-compose.yml

services:
  sapixdb:
    image: sapixdb/agent:latest
    container_name: sapixdb
    restart: always
    ports:
      - "7475:7475"
    volumes:
      - /var/lib/sapixdb/strand:/data/strand
      - /var/lib/sapixdb/graph:/data/graph
      - /var/lib/sapixdb/blobs:/data/blobs
    environment:
      SAPIX_AGENT_ID:         production-primary
      SAPIX_MASTER_SEED:      ${SAPIX_MASTER_SEED}
      SAPIX_KEYPAIR_SEED_HEX: ${SAPIX_KEYPAIR_SEED_HEX}
      SAPIX_ROOT_KEY:         ${SAPIX_ROOT_KEY}
      SAPIX_LICENSE_KEY:      ${SAPIX_LICENSE_KEY}
      SAPIX_STRAND_DIR:       /data/strand
      SAPIX_GRAPH_DIR:        /data/graph
      SAPIX_BLOB_DIR:         /data/blobs
      SAPIX_BIND_ADDR:        0.0.0.0:7475
      SAPIX_CODIOS_URL:       ${SAPIX_CODIOS_URL}
      SAPIX_CODIOS_API_KEY:   ${SAPIX_CODIOS_API_KEY}
    healthcheck:
      test: ["CMD", "curl", "-f", "http://localhost:7475/v1/health"]
      interval: 30s
      timeout:  10s
      retries:  3

Production secrets management

Never put secrets in docker-compose.yml directly. Use: - Railway: shared environment variables - AWS: Secrets Manager + ECS task role - Kubernetes: sealed secrets or external secrets operator - Docker Swarm: docker secret

# Pass from environment
SAPIX_MASTER_SEED=<hex> SAPIX_ROOT_KEY=<key> docker compose up -d

Mesh replication (high availability)

Register a peer node:

curl -s -X POST http://localhost:7475/v1/mesh/peers \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer spx_root_YOUR_ROOT_KEY" \
  -d '{
    "peer_id":  "replica-01",
    "endpoint": "http://replica-host:7475",
    "api_key":  "spx_root_REPLICA_ROOT_KEY"
  }' | python3 -m json.tool

SapixDB pushes all writes to registered peers. Both nodes share the same SAPIX_MASTER_SEED (same encryption key, same data). Reads can go to either node.

Production checklist

Performance tuning

Use batch writes for all bulk ingestion (Lesson 8). Create indexes on frequently filtered fields (Lessons 18–20). Use count and aggregate instead of loading full records for stats.

Final challenge

Deploy a two-node SapixDB setup with mesh replication. Write a record to node 1. Verify it appears on node 2. Run chain verification on both nodes.

---

## Summary — All 50 Lessons

---

*Lessons 1–6 complete. Proceed from Lesson 7 onward one at a time.*