Render 502 Bad Gateway Error

How to fix it

Try these in order — the most common fix is first.

1. 1. Check the deploy and logs

   Copy steps
   1. Confirm the latest deploy is Live
   2. Read runtime logs around the time of the error
   Did this fix it? Yes No

2. 2. Fix host, port and health check

   Copy steps
   1. Bind to 0.0.0.0:$PORT
   2. Make the health check path return 2xx or 3xx quickly
   Did this fix it? Yes No

3. 3. Fix intermittent 502 on Node

   1. Increase keepAliveTimeout and headersTimeout above 120 seconds on your HTTP server
   Did this fix it? Yes No

4. 4. Fix worker timeouts on Python

   1. Raise the Gunicorn --timeout value or speed up slow requests
   Did this fix it? Yes No

Symptoms

Visiting the service returns 502 Bad Gateway, sometimes only intermittently.

Why it happens

• No successful deploy yet
• Instance failing health checks
• Server not bound to 0.0.0.0 and PORT
• New custom domain still propagating (up to an hour)
• Node keepAliveTimeout lower than the proxy's, giving Connection reset by peer
• Gunicorn worker timeout (WORKER TIMEOUT, SIGKILL)

This guide also covers

• render 502 bad gateway
• render 503 service unavailable
• render connection reset by peer

The same problem in other apps

All export, render & quality loss guides →

• Railway Slow Deployments and Slow App (Railway)
• Fly.io Deploy Failed: Waiting for Depot Builder, Registry 401, Too Many Requests (Fly.io)
• Vercel Deployment Failed: Build Error (Vercel)
• DigitalOcean App Platform Build Failed (DigitalOcean)
• Heroku Build Failed Error (Heroku)

Other Render errors

• Render Free Instance Spins Down: Slow First Request
• Render "No Open Ports Detected" Deploy Error
• Render Build or Service Out of Memory
• Render Build Failed: Exited With Status 1
• Render Free Postgres Expired or Deleted
• Render Custom Domain Verification or SSL Pending
• Render Health Check Failed or Deploy Timed Out
• Render Free Service Suspended: 750 Hours Used Up