I have long running jobs on different services. I instrumented the solution to trace these jobs. They can run for hours and perform a large number of Elasticsearch requests. I would like to know what is going on when jobs start failing. Ideally I would like to create a span only when requests return an http error code or when the request time is beyond some configurable value.
At the moment, kibana cannot handle that much spans in one transaction. (Is it possible to reduce the time frame to make spans appear?) When testing this with jaeger I could display such a large amount of requests.
Is it possible to be more fine grained?
I use .net 6 to develop our services.