Open slarwise opened 2 years ago
@slarwise can't remember exactly what this is about. Can you share a print screen or something.
Do I understand you correctly that each time the health endpoint of ingress-nginx is queries datadog generates a traces?
NGINX DataDog specific tracing is deprecated in favor of OTEL.
This tracing results in around 200-300 spans per second for posting metrics, which results in a lot of noise. It is useful to have tracing enabled for nginx when the applications are used, since services might communicate with each other through nginx, and then we want the trace information to be propagated over nginx. Ideally we would disable tracing for the metrics endpoint only.