Introducing JavaScript Exception Severity Scoring in Embrace Web RUM

Traditional error dashboards treat every exception as equal. You might be able to sort a list of exceptions by frequency, or even number of affected users, but that has typically been the extent of it. 

This is limiting for real impact analysis because, in practice, certain errors hold much greater weight when it comes to the actual functionality they disrupt. A single uncaught error in your checkout flow, for example, is far more critical than dozens of minor console warnings in a background script. Without prioritization, engineering teams spend too much time triaging noise and not enough time solving meaningful problems.

Severity Scoring eliminates that guesswork. By automatically analyzing exception patterns, affected sessions, and downstream effects, Embrace assigns each exception a severity score that reflects its user impact.

When Embrace detects a JavaScript exception, our algorithm takes into account a number of variables to calculate its ultimate severity score. 

This begins with the base severity of the exception, which depends on whether the exception was handled (lower base severity number) or unhandled (higher base severity number). Unhandled exceptions receive a higher base score because they typically indicate more serious problems that could cause the application to stop functioning or lead to degraded user experience.

Then, the algorithm factors in a few contextual multipliers to refine the severity score based on certain conditions related to the exception. 

Some examples of contextual multipliers that Embrace factors into calculating the severity score address variables like: 

• Whether the exception occurred in a third-party library vs. in the application code
• How quickly the session ended after the exception
• Whether an exception occurred right after a user action
• Whether an exception was in close proximity to a network failure 

For a complete list of all contextual multipliers and greater detail into how Embrace calculates severity score, check out our docs. 

In the Embrace dashboard, exceptions are grouped by their stack trace line and displayed as a list. Each row represents a group of exception events that share the same location in an application’s code.

When looking at the Severity Score, which is now represented as a new column in the JS Exceptions page, this score is actually the average Severity Score across all exception events in that group. 

This average score helps teams quickly identify which exception groups are causing the most severe issues for their users.

Additional columns provide context for prioritization:

• Events: Total count of exception occurrences in this group
• Devices: Unique devices affected by this exception
• Users: Proportion of your user base experiencing this exception
• Versions: Version range where this exception has been observed
  

By combining the average severity score with these metrics, teams can prioritize exceptions that are both severe and widespread. For example, an exception with a high severity score affecting a large percentage of users should be addressed before a similarly severe exception affecting only a few devices.

Severity Scoring is built to work seamlessly with your existing Embrace workflows:

• Exception dashboards now highlight the most severe issues first.
• Filtering and search make it easy to isolate critical or recurring errors.
• Session-level analysis links each exception directly to the user experience it impacted.

No more guesswork. No more wasted triage cycles. Just a clear hierarchy of what’s breaking your web experience, why, and what directions to take to resolve it. 

JavaScript Exception Severity Scoring is now live for all Web RUM customers. Start using it today to identify, prioritize, and fix the errors that matter most to your users.

Not an Embrace customer yet? Sign up for free today and start tracking and analyzing your exceptions in the dashboard.