CPU can be at 20 % while a specific request takes eight seconds because it fired thirty database queries. The Slow Request Log catches that class of problem: it records web requests that exceeded a duration threshold, along with the queries and hotspots inside them.

Note

Available on the Business plan and above.

What you see

Under Slow Requests in the sidebar. The index is a table of slow requests across your monitored applications:

• Timestamp.
• Server / application.
• HTTP method and route.
• Duration (with a visual indicator vs. threshold).
• Response status.
• Query count and total DB time.

Filters: server, route, status, time window, duration threshold.

Click a row to open the detail view:

• Full request metadata (headers, params, user if authenticated).
• Query list — every SQL statement that ran, ordered by duration, with row count and time spent.
• Timing breakdown — application, database, and other.
• Repeated-query detection (N+1 offenders bubble up).

How it works

• A small piece of application middleware measures each request end to end. When the total duration exceeds the configured threshold, the request is recorded along with a summary of every query it issued.
• Only slow requests are stored — normal traffic passes with no overhead beyond the timer.
• Recording is opt-in per application: you enable the middleware and set the threshold.
• Query text is stored so you can see which statement is expensive. Bound parameter values are captured in the same form your application logs them.

What you can do

• Filter and sort the index to find the worst offenders or a specific route.
• Open a request to see the query list and figure out whether the fix is an index, a cache, or a rewrite.
• Cross-link to the server’s overall metrics for the same window — often the slow request coincides with a load spike.

Related

• Diagnostic Captures — on-demand CPU forensics for the “the whole box is slow” case.
• Activity — slow-request events appear on the unified timeline.