Deployment for Testing and Production
Production Deployment
Deployment for Testing and Production
Production Deployment
Requirements
This requirements can support up to
3000 active users daily. If you have more users, you can scale the resources as needed.
- Kubernetes Cluster
- 3 Master Nodes: Each one 4 CPU, 16 GB RAM
- 6 Worker Nodes: Each one 8 CPU, 32 GB RAM
- Persistence Storage
- s3, amazon-s3, azure-blob, aliyun-oss, google-storage, tencent-cos
- 512 GB Storage
- Postgres Database
- Version 14+
- 4 CPU, 8 GB RAM
- 512 GB
SSDStorage
- Redis Database
- Version 6+
- 2 GB RAM
- Vector DB Virtual Machine - Qdrant Cluster
- 3 Virtual Machines
- 8 CPU, 64 GB RAM
- 100 GB Storage
Dify Services Resources
| Services | Replicas | CPU | RAM - GB | Total CPU | Total RAM - GB |
|---|---|---|---|---|---|
| dify-api | 3 | 4 | 10 | 12 | 30 |
| dify-worker | 3 | 3 | 10 | 9 | 30 |
| unstructured | 3 | 1 | 4 | 3 | 12 |
| dify-web | 3 | 2 | 8 | 6 | 24 |
| dify-sandbox | 3 | 1 | 4 | 3 | 12 |
| ssrf-proxy | 3 | 1 | 4 | 3 | 12 |
| dify-enterprise | 3 | 1 | 4 | 3 | 12 |
| dify-enterprise-frontend | 3 | 1 | 4 | 3 | 12 |
| Total | 42 | 144 |
Deployment
Initial External Resources
- Persistence Storage
- Postgres Database
- Redis Database
- Vector DB Virtual Machine
Update Helm Chart Values
Update the values-no-infra-for-prod.yaml file.
- Update
persistencewith your persistence storage configuration - Update
externalPostgreswith your postgres database configuration - Update
externalRediswith your redis database configuration - Update
vectorDBwith your vector db virtual machine configuration
Install Dify Enterprise
You can install Dify Enterprise using the following guide. But you need to use values-no-infra-for-prod.yaml instead of values.yaml.
Verify Services Status
After the installation, you can verify the services status referring to the following guide.