Overview
The profiling engine helps diagnose runtime performance issues. It is an advanced diagnostic feature rather than a day-to-day configuration item.
When to use it
Use profiling when CPU usage, memory usage or routing latency is abnormal and ordinary logs are not enough to identify the cause.
Workflow
Reproduce the issue with a minimal profile, collect a short profile window, then compare the result with recent configuration changes or rule-provider updates.
Cautions
Profiling output can include operational details. Share it carefully and remove sensitive endpoints or subscription information before posting it publicly.
Related pages
Reference examples
These examples mirror the corresponding Chinese documentation page so the English page carries the same configuration material.
profile:
tracing: true