When trying to access the UI of an f5 virtual appliance (hosted on ec2) via Secure Access Cloud, you discover that not all of the objects load on one of the pages. The objects which fail to load are part of a wider dashboard needed for monitoring purposes.
{ "title": "Application unreachable",
"message": "You cannot access the application (HTTP status 526).",
"reason_message": "We failed processing the response",
"remediation_message": "Please contact your administrator for assistance.",
"additional_details": "Failed Processing Response: Failed processing HTTP Response 'HTTP/1.1 200 OK' (request-id: 1234, application-id: #########, connector-id: ##########).. Exception: null",
"code": 526,
"reason_phrase": "Failed Processing Response",
"show_error": true,
"request_id": "####################",
"template_file": "branded_error_dude.html",
"show_contact": true,
"show_support_contact": false,
"username": "user.name@example.com"
}
SAC Proxy not handling compressed response from Web server correctly.
Fixed in the SAC code update for April '23. When these issues occur with compressed payloads that SAC cannot understand, the following workarounds may be possible:
HAR file decodes the responses to display, and hence you cannot get the info you need to replicate easily.
Using curl, you can generate the request to the Web server directly and save the output with -o parameter:
curl 'https://xx.xx.xx.xxx/tmui/tmui/dashboard/app/main/dashboardManager.template.html' -o response.gz
Now that you have the output file, you can add this to an internal testing web server and replicate the issue easily.