How to optimize function memory and duration on Vercel

Create or update your vercel.json file in the project root. Add memory configuration for specific functions:

{
  "functions": {
    "api/heavy-function.js": {
      "memory": 1024
    },
    "api/light-function.js": {
      "memory": 128
    }
  }
}

Set memory values between 128MB and 3008MB based on your function's requirements.

Configure the maxDuration property in your vercel.json to prevent functions from running indefinitely:

{
  "functions": {
    "api/*.js": {
      "maxDuration": 10
    }
  }
}

Duration is set in seconds. Free plans are limited to 10 seconds, Pro plans up to 60 seconds, and Enterprise up to 900 seconds.

Implement memory-efficient coding practices:

• Use const and let instead of var for proper garbage collection
• Avoid loading large dependencies unnecessarily
• Process data in chunks rather than loading everything into memory
• Use streaming for large file operations
• Clear variables when done: largeObject = null

Reuse database connections and cache frequently accessed data:

// Global connection variable
let cachedConnection = null;

export default async function handler(req, res) {
  if (!cachedConnection) {
    cachedConnection = await createConnection();
  }
  // Use cached connection
}

Use external caching services like Redis or Vercel's Edge Config for shared state.

Access the Functions tab in your Vercel dashboard to monitor:

• Memory usage patterns
• Execution duration
• Cold start frequency
• Error rates and timeouts

Click on individual function invocations to see detailed metrics and logs. Use this data to identify optimization opportunities.

Reduce cold start penalties by:

• Keeping function code lightweight
• Using Vercel's Edge Functions for simple operations
• Implementing connection warming strategies
• Avoiding heavy initialization code
• Using import statements at the top level only for critical dependencies

Deploy your optimized functions using vercel --prod or through Git integration. Test performance improvements by:

• Running load tests on optimized endpoints
• Comparing before/after metrics in the dashboard
• Monitoring memory usage under different loads
• Verifying execution times meet requirements