Users experienced connection failures when attempting to access websites through Cisco Secure Access using the SWG (Secure Web Gateway) proxy URL swg-url-proxy-https-8xxxx.sseproxy.qq.opendns.com.
The specific symptoms included:
Web requests (such as to https://www.google.com) were failing with connection errors in browsers
Traffic was not appearing in activity search logs
Server resets were observed from the client source IP/Client Egress IP toward destination endpoints:
k8s-sigpro-sigpro-c10eb6eb-38fbeb0455191ac5.elb.eu-central-1.amazonaws.com
k8s-sigpro-sigpro-f8f3d861-14341dd91e659675.elb.eu-central-1.amazonaws.com
Issues began at approximately 09:30 AM CEST
DNS resolution for the SWG hostname was functioning correctly
Telnet connections to swg-url-proxy-https-8xxxx.sseproxy.qq.opendns.com on port 443 were successful
The outer firewall was not blocking connections to the destination IPs
Packet capture analysis revealed TCP RST packets coming from the proxy, indicating connection termination at the proxy level.
Technology: Cisco Secure Access (SSE)
Component: Secure Web Gateway (SWG)
SWG Proxy URL: swg-url-proxy-https-8385532.sseproxy.qq.opendns.com
Affected Ports: 443 and 80
Original NAT IP: 20.x.x.x
Updated NAT IP: 21.x.x.x
AWS ELB endpoints in for ex: eu-central or us-east region
WPAD configuration directing browsers to SWG proxy
The issue was resolved by updating the allow-list configuration to include the correct NAT IP address.
The steps described in the next sections were taken to identify and resolve the problem.
Collected packet captures and WPAD script configuration to analyze the traffic flow and proxy configuration.
In addition to that, main clue was https://policy.test.sse.cisco.com was showing "401 unauthorized" error. Which does meant http connect request is hitting to proxy, but proxy not able to authenticate user traffic based on their OrgID and other meta data details to apply policy and allow the traffic.
Analysis of the packet capture revealed:
TCP RST packets were being sent from the proxy
Activity search logs did not show the failing traffic
The source IP in the traffic flow had changed
Investigation revealed that the NAT public IP address had changed from 20.x.x.x to 21.x.x.x. Changed IP was added as registered network in the user CSA UI, SWG traffic started to work for PAC file deployment after registering correct IP in the CSA portal.
Updated the allow-list/firewall rules to permit traffic from the new NAT IP address 21.x.x.x.
After updating the allow-list with the new NAT IP address, normal Secure Access traffic flow was restored and web requests began functioning correctly through the SWG proxy.
The root cause was a change in the user NAT public IP address from 20.x.x.x to 21.x.x.x. The Secure Access SWG proxy was configured to only allow traffic from the original IP address, causing connection resets when traffic arrived from the new IP address. In addition to that, main clue was https://policy.test.sse.cisco.com was showing "401 unauthorized" error, which refers to http connect request is hitting to proxy, this is confirmed by PCAP as well, but proxy not able to authenticate user traffic based on their OrgID and other meta data details to apply policy and allow the traffic. This resulted in the proxy terminating connections with TCP RST packets, preventing successful web requests from being processed.
Added the new NATed IP in user CSA UI under registered network to resolve web traffic flow issue for SWG PAC file.
| Revision | Publish Date | Comments |
|---|---|---|
1.0 |
17-Aug-2026
|
Initial Release |