Skip to content

TROUBLESHOOTING.md

Latest commit

 

History

History
185 lines (134 loc) · 4.62 KB

TROUBLESHOOTING.md

File metadata and controls

185 lines (134 loc) · 4.62 KB

🔧 QUICK FIX GUIDE - Backend Connection Issues

Problem: Variations Just Buffer/Load Forever

✅ SOLUTION STEPS (Follow in order):

Step 1: Wake Up Your Render Backend (MOST COMMON ISSUE)

Render's free tier sleeps after 15 minutes of inactivity. It takes 30-60 seconds to wake up on first request.

Test if backend is awake:

# Open this in your browser or use curl
https://retexture.onrender.com/health

Expected Response:

{"status":"healthy"}

If you get 502 Bad Gateway:

  • ✅ This is NORMAL - service is waking up
  • ✅ Wait 30-60 seconds
  • ✅ Try again
  • ✅ Once it responds, try variations again

Pro Tip: Keep this tab open to ping every few minutes:

<!-- Save as wake-render.html and open in browser -->
<script>
setInterval(() => {
  fetch('https://retexture.onrender.com/health')
    .then(r => r.json())
    .then(d => console.log('✓ Backend alive:', d))
    .catch(e => console.log('✗ Waking up...', e));
}, 300000); // Every 5 min
</script>

Step 2: Verify Environment Variable in Vercel

  1. Go to https://vercel.com/dashboard
  2. Select your ReTexture project
  3. Go to Settings → Environment Variables
  4. Check NEXT_PUBLIC_API_URL exists and equals: https://retexture.onrender.com
  5. If missing or wrong, add/fix it
  6. Important: Go to Deployments → Click ••• → Redeploy

Step 3: Check Render Logs

  1. Go to https://dashboard.render.com
  2. Click your retexture-backend service
  3. Click Logs tab
  4. Look for errors when you try to generate variations

Common issues in logs:

  • GOOGLE_API_KEY not found → Add it in Render environment variables
  • ModuleNotFoundError → Build may have failed, redeploy
  • Port already in use → Render issue, redeploy service

Step 4: Test Backend Locally

Open the test file I created: backend-test.html in your browser

# Or test with curl
curl https://retexture.onrender.com/health
curl https://retexture.onrender.com/

Step 5: Check Browser Console

  1. Open your Vercel site: https://retexture.vercel.app
  2. Open Developer Tools (F12)
  3. Go to Console tab
  4. Try to generate variations
  5. Look for errors

Common errors:

❌ "Failed to fetch"

  • Backend is down or sleeping
  • Wrong backend URL
  • Solution: Wait for backend to wake up

❌ "CORS error"

  • Backend CORS not configured
  • Solution: Check app/main.py has CORS middleware (already added)

❌ "502 Bad Gateway"

  • Backend is cold starting
  • Solution: Wait 30-60 seconds and retry

❌ "Request timeout"

  • Backend taking too long
  • Solution: Check Render logs, may need to upgrade from free tier

🚀 QUICK CHECKLIST:

  • Backend responds at https://retexture.onrender.com/health
  • Vercel has NEXT_PUBLIC_API_URL=https://retexture.onrender.com
  • Vercel redeployed after adding env variable
  • Render has GOOGLE_API_KEY set
  • Waited 30-60 seconds for cold start on first request
  • Browser console shows no errors

💡 STILL NOT WORKING?

Test API directly:

# Test variations endpoint (replace with actual base64 image)
curl -X POST https://retexture.onrender.com/generate/variations/stream \
  -H "Content-Type: application/json" \
  -d '{"image_data":"BASE64_IMAGE_HERE","concept":"product photography"}'

Check environment variables are loaded:

In Vercel (Frontend):

// Add this to any page temporarily
console.log('Backend URL:', process.env.NEXT_PUBLIC_API_URL);
// Should show: https://retexture.onrender.com

In Render (Backend):

  • Check Render dashboard → Environment → Variables
  • Should have GOOGLE_API_KEY

🎯 MOST LIKELY CAUSES (in order):

  1. 75% chance: Render free tier sleeping (wait 30-60 sec)
  2. 15% chance: Missing/wrong NEXT_PUBLIC_API_URL in Vercel
  3. 5% chance: Missing GOOGLE_API_KEY in Render
  4. 5% chance: Vercel not redeployed after env variable change

📊 How to Know It's Working:

✅ Backend responds to /health in < 2 seconds ✅ Frontend shows "Starting AI service..." toast ✅ After 5-10 seconds, first variation appears ✅ All 3 variations generated within 30-45 seconds

💰 Upgrade to Fix Forever:

Problem: Free tier sleeps = slow Solution: Upgrade to Render Starter ($7/month)

Benefits:

  • ⚡ Always running (no cold starts)
  • 🚀 Instant responses
  • 💪 Better performance
  • 📈 More RAM (512MB)

To upgrade:

  1. Go to Render dashboard
  2. Select your service
  3. Click Upgrade → Starter
  4. Confirm

Need more help? Check backend logs or DM me!