Azure 502 Bad Gateway Issue

Hi @Mohammed Shafi

Welcome to the Microsoft Q&A Platform! Thank you for asking your question here.

Please go through the below points:

• Your VM might be intermittently overloaded (like high CPU, memory, or network usage), making it unable to respond to requests from the application gateway health probes or user traffic in a timely manner.
• Please use Azure Monitor to track CPU, memory, network, and disk I/O on your VM. Look for spikes or sustained high usage around the times when the 502 errors occur. For your reference: https://zcusa.951200.xyz/en-us/azure/virtual-machines/monitor-vm#:~:text=Simplified%20onboarding%20of%20the%20Azure%20Monitor%20agent%20and%20the%20Dependency%20agent%2C%20so%20that%20you%20can%20monitor%20a%20virtual%20machine%20(VM)%20guest%20operating%20system%20and%20workloads.
• The web application running on your VM might be restarting intermittently, causing temporary unavailability. Please try to review your web application's logs for errors or exceptions that might indicate crashes.
• The backend server closes connections, but the application gateway still sends traffic, make sure that the backend server's keep-alive timeout is greater than application gateway idle timeout.
• Try to review the NSG rules associated with both your application gateway subnet and your VM's subnet. Ensure there are rules allowing inbound traffic on the ports used by your application (like 80 or 443) from the application gateway IP range.
• Verify that your UDRs are not directing traffic away from your VM. Ensure that traffic destined for your VM's IP address is routed correctly.
• If you're using the VM's firewall, ensure that there's a rule allowing inbound traffic on the required ports from the application gateway IP address range
• Ensure the health probe interval is not too short (30 sec is common). A very short interval can overwhelm a busy server.
• Make sure the timeout is appropriate (like 30 sec). It should be long enough for the server to respond even under load.
• An appropriate threshold prevents flapping between healthy and unhealthy states. Verify the port used in your backend settings matches the port your application is listening on within the VM.
• If you are enabled the connection draining, be sure to configure an appropriate timeout for connection draining.
• Analyze your application gateway WAF logs to see if any requests are being blocked around the time of the 502 errors.
• Make sure that your DNS records are correctly configured and that the DNS servers you're using are reliable.
• As you mentioned the issue resolves itself or with a browser cache clear, it might be a temporary network glitches or routing problems within Azure's infrastructure could be causing the connection failures.
• Enable diagnostic logging for your application gateway and send the logs to a Log Analytics workspace. This will give you detailed insights into health probe failures, request routing, and WAF activity.
• Use Network Watcher's connection troubleshoot feature to diagnose network connectivity issues between your application gateway and VM. For your reference: https://zcusa.951200.xyz/en-us/azure/network-watcher/connection-troubleshoot-portal#:~:text=In%20this%20article%2C%20you%20learn%20how%20to%20use%20the%20connection%20troubleshoot%20feature%20of%20Azure%20Network%20Watcher%20to%20diagnose%20and%20troubleshoot%20connectivity%20issues.%20For%20more%20information%20about%20connection%20troubleshoot%2C%20see%20Connection%20troubleshoot%20overview.

Kindly let us know if the above helps or you need further assistance on this issue.

Thanks,

Sai.