Troubleshooting request ECONNRESET errors

Understand what ECONNRESET means in Vercel runtime logs, why it happens when calling external APIs, how to diagnose it, and when to contact Vercel Support.

Guides/Functions
2 min read
Last updated October 15, 2025

ECONNRESET means connection reset by peer. The remote server (or something on the network path) closed the connection unexpectedly. This is usually caused by the external service or a misconfiguration, not by Vercel.

  • Persistent errors: Usually a setup or configuration issue (wrong protocol, port, or missing allowlist).
  • Intermittent errors: Often external (e.g., provider maintenance, load spikes, or deployment changes). Check provider logs and metrics to correlate timing.

Other possible triggers:

  • Provider timeouts or rate limiting
  • Protocol/TLS mismatches
  • IP allowlisting issues (default Vercel egress IPs are dynamic; use Secure Compute or Static IPs for dedicated IPs)
  • Running functions far from the provider region (learn to configure your Vercel Function Region here).

  1. Check with your provider: Did they receive the request? (If yes, proceed to contact Vercel Support)
  2. Inspect Vercel Runtime Logs: Look for request details and error timing.
  3. Enable NODE_DEBUG: Add an environment variable to your project NODE_DEBUG=http,https,net to capture detailed network activity.
  4. Add timeouts and retries: Handle transient resets with capped backoff retries.
  5. Limit concurrency: Avoid flooding the provider with parallel requests.

If your provider confirms they received the request but you still see ECONNRESET, reach out via the Support Center. Include:

  • Deployment URL and function route
  • Timestamp (UTC) and function region
  • Runtime (Node.js)
  • Runtime Logs for the request
  • Any additional information that the provider can provide about the request

  • Place your functions in regions close to your data source
  • Use Secure Compute or Static IPs for dedicated egress IPs if your provider requires allowlisting
  • Always set timeouts, retries, and concurrency limits
  • Monitor patterns with Observability to spot issues early

Was this helpful?

supported.