Deploying Flowise Apps to Production: Handling High Request Volumes

🔑 Henry explains how to deploy Flowise to cloud services and allow multiple users to connect to a single Shuffle of Flowise.

🧪 Henry introduces a new feature called load testing, which allows stress testing of Flowise APIs.

💻 Henry demonstrates the deployment process using DigitalOcean and the need to persist flows to avoid data loss.

🔑 Using SSH key is recommended for secure access to the droplet.

🖥️ The basic plan on DigitalOcean can handle 100-1000 requests per second without any problems.

📈 Auto scaling on AWS allows handling more requests by spinning up additional instances.

🚀 Deploy Flowise apps to production and handle high request rates.

💾 Digital Ocean provides built-in storage, eliminating the need for additional purchases.

🔒 Render addresses the persistent storage issue by allowing users to purchase built-in memory.

🐳 Using Docker is the recommended installation method for Flowise.

📝 Set up Flowise by installing Docker, Docker Compose, and modifying environmental variables.

💻 Render is a user-friendly option for quick setup, while Digital Ocean offers more control over the instance.

🚀 You can deploy Flowise apps to production on DigitalOcean by running Docker Compose commands.

🔄 To update your Flowise app to the latest release, stop the applications, pull the latest image, and restart the app using Docker Compose.

🔒 To secure your app and configure a custom domain, you need to install and configure Nginx, set up a reverse proxy, and configure DNS records.

🔍 Reverse proxy setup is done within the party window of the virtual machine.

🔒 Authentication can be added to Flowise by specifying username and password in the environmental variables.

💾 Options for vector stores on DigitalOcean include using files, Voice, and Super Bass.

📁 Adding a folder with files module is possible on DigitalOcean, Railway, and Render.

🔑 Deploying Flowise apps to production requires specifying the path from the virtual machines on Digital Ocean.

📂 Folders with files in Flowise support specific formats like PDF and JSON, and only the supported file types can be loaded.

💻 Exporting JavaScript code from Flowise is still in progress and will be implemented once more Integrations are added.

💡 You can export and import databases, including flows, messages, credentials, and API keys, to handle 100-1000 requests per second for live apps.

✅ Increasing the top K size and using metadata can improve the performance of the AI response when retrieving embeddings.

🔗 There are plans to parameterize API endpoints and open them up for integration with other internet APIs.