Handling Parent-less Spans in Batch Request Resolvers
One source of parent-less spans, aka transactions, remain, which is batch request resolvers, it appears to generate the highest count of transaction and therefore usage e.g at Sentry.
Now I could disable sampling for them. But I wonder if anyone else has a different strategy in mind for that.
The main conundrum is the nature of batch resolvers with cache; nobody really owns these operations.
Though on some level, I think it would be fine if they get attributed to the span who caused a miss.
Now I could disable sampling for them. But I wonder if anyone else has a different strategy in mind for that.
The main conundrum is the nature of batch resolvers with cache; nobody really owns these operations.
Though on some level, I think it would be fine if they get attributed to the span who caused a miss.
