Debug commands for Cased CD Enterprise 502 Bad Gateway error

debug-502-commands.md

Debug Commands for Cased CD Enterprise 502 Error

Run these commands inside the pod to diagnose the 502 Bad Gateway issue.

Priority 1: Check Backend Status

# 1. Check if the backend process is running
ps aux | grep cased-backend

Expected output: Should show /app/cased-backend process running

# 2. Test backend directly (should return "OK")
curl -v http://127.0.0.1:8081/health

Expected output: OK

# 3. Check what port the backend is listening on
netstat -tlnp | grep 8081

Expected output: Should show 0.0.0.0:8081 or 127.0.0.1:8081

Priority 2: Check Nginx Configuration

# 4. Check nginx error logs for more details
tail -50 /var/log/nginx/error.log

Look for errors related to:

• DNS resolution failures
• Connection refused
• Upstream timeout

# 5. Check what's the proxy_target in nginx config
cat /tmp/nginx.conf | grep -A2 "set \$proxy_target"

Expected output: set $proxy_target "http://localhost:8081";

# 6. Check DNS resolver in nginx config
cat /tmp/nginx.conf | grep resolver

Expected output: resolver <IP> valid=10s ipv6=off;

Priority 3: Test DNS and Network

# 7. Test if nginx can resolve localhost via DNS
nslookup localhost

Important: This will likely fail! Nginx resolver only queries DNS, not /etc/hosts

# 8. Check if backend is accessible on localhost
curl -v http://localhost:8081/health

Expected output: OK

# 9. Check environment variables
env | grep -E '(BACKEND_PORT|PROXY_TARGET|DNS_RESOLVER)'

Expected output:

BACKEND_PORT=8081
PROXY_TARGET=http://localhost:8081
DNS_RESOLVER=<IP>

Priority 4: Check Logs

# 10. Check recent nginx access logs for API requests
tail -100 /var/log/nginx/access.log | grep "api/v1/settings"

Look for 502 errors

Quick Test: Direct Backend Access

# Try accessing backend API endpoint directly
curl http://127.0.0.1:8081/api/v1/settings/audit

If this works but nginx proxy doesn't, it's definitely a proxy configuration issue.

What to Look For

If backend is NOT running:

• Check pod logs: kubectl logs <pod-name> -n argocd
• Look for backend startup errors
• Check if backend binary exists: ls -la /app/cased-backend

If backend IS running but not accessible:

• Check if listening on correct interface (should be 0.0.0.0:8081)
• Check firewall rules (unlikely in container)

If nginx can't reach backend:

• Check nginx error logs for "upstream" errors
• Check if localhost resolves (it probably won't via DNS)
• Try using 127.0.0.1 instead in nginx config

Expected Root Cause

Based on the symptoms, the most likely issue is:

Nginx resolver can't find localhost because:

1. Nginx's resolver directive only queries DNS servers
2. DNS servers don't have a record for "localhost"
3. Even though /etc/hosts has 127.0.0.1 localhost, nginx's resolver ignores it

Solution: Use 127.0.0.1 directly instead of localhost in the nginx proxy configuration.