What can I do about Vercel Functions timing out?

Vercel Functions can time out from long-running tasks, inefficient code, or upstream service issues.

The best way to avoid hitting these limits is by taking advantage of Fluid Compute, which blends serverless flexibility with server-like capabilities, including running functions up to 1 minute on free plans and 14 minutes on paid plans.

Fluid Compute offers a hybrid solution that removes many of the traditional limitations of serverless functions:

• Optimized concurrency: Multiple function invocations can run on a single instance, reducing cold starts and improving resource utilization.
• Extended durations: Enjoy longer maximum durations (up to 800 seconds on Pro and Enterprise) without additional configurations.
• Dynamic scaling & automatic cold start optimizations: Scale seamlessly during traffic spikes and benefit from bytecode caching to reduce cold starts.
• Background processing: Use waitUntil (or after with Next.js) to continue tasks after sending an initial response.

If you frequently encounter timeouts or want to run network-intensive tasks, enabling Fluid Compute is your most effective long-term solution.

• Open your project in the Vercel dashboard
• Click on Settings and select the Functions section
• Scroll to the Fluid Compute section and enable the toggle for Fluid Compute
• Redeploy your project to apply changes.

Fluid Compute is currently supported by the Node.js and Python runtimes.

• By handling multiple invocations within a single function instance, you spend less time waiting for cold starts and better utilize idle compute resources.
• This is especially helpful for I/O-bound workloads (e.g., calling external APIs or databases).

• Fluid Compute has higher default and maximum duration limits than traditional serverless functions, making it less likely you’ll hit a timeout for long-running tasks.

If you’re still encountering timeouts even after enabling Fluid Compute, review the following common causes:

• Check API or database call durations: If these calls exceed your default or configured duration, you’ll see timeouts.
• Extend your function’s duration if needed: You can override Fluid Compute defaults if needed for even longer running functions. If building AI applications, we recommend streaming.

• Always send an HTTP response (even an error). If your function never returns anything, Vercel will wait until the maximum duration has elapsed and then time out.

• Inspect your logic for any loops or recursive calls that never terminate. This is a common cause of unintended long-running functions. Inspect your function runtime logs and observability.

• Verify third-party integrations or database connections. If upstream services fail to respond, make sure you handle the error gracefully and return a response (e.g., an error message) rather than waiting indefinitely.