What cross-signal correlation is #
Cross-signal correlation is the process Loupely uses to connect browser-side signals with server-side signals and identify the causal relationship between them. It’s the core of what makes Loupely’s diagnosis more useful than looking at error logs alone.
A browser-side signal is something the browser observed: a JavaScript error, a network request that returned a 500 status code, a failed AJAX call, a console warning. A server-side signal is something that happened on the server: a PHP Fatal Error, a hook that didn’t fire, a database write that failed, a WooCommerce pipeline stage that didn’t complete. Each side captures a partial view of the failure. Cross-signal correlation finds the connection between them.
Why you need both sides #
Consider a checkout that fails silently. The customer clicks Place Order. Nothing happens. No error message. From the browser’s perspective: a JavaScript Ajax Request fired, it received a 500 response, the JavaScript error handler ran, the Checkout Button reset to its default state. From the server’s perspective: a PHP Fatal Error fired inside the Payment Gateway plugin when it tried to process the charge, the script stopped executing, PHP returned an empty 500 response to the browser.
Looking at the browser side alone tells you there was a JavaScript error and a 500 response. It doesn’t tell you what caused the 500. Looking at the PHP error log alone tells you there was a fatal error in the Payment Gateway plugin. It doesn’t tell you that this specific error is what broke this specific checkout attempt at this specific timestamp. Cross-signal correlation connects the two: the PHP fatal at timestamp X caused the 500 response that triggered the JavaScript error that caused the Checkout Button to do nothing.
How the correlation rules work in practice #
Correlation rules are patterns that define: which browser-side signal type, which server-side signal type, and which timing relationship between them constitute a known failure class. When a capture matches a rule precisely, Loupely assigns a high-confidence diagnosis naming the cause. When signals are present but don’t match a known pattern exactly, confidence is lower. When signals don’t match any rule, the result is null.
The rule set grows as new failure patterns are identified. If you encounter a failure that produces consistent signals but Loupely doesn’t diagnose, that pattern is a candidate for a new correlation rule. Contact support@useloupely.com with the capture file, and the team can evaluate whether the signals represent a diagnosable pattern. See Correlation Rules: How Loupely Connects Browser and Server Signals for more detail on how the rules are structured.