Why not have a trace for every job?
Opened this issue · 2 comments
Very strange setup here - there is a single trace for the entire queue processor, and each job processed is merely a span in that huuuge trace:
Why not have a trace for every job instead? We want to be able to figure out which particular jobs are slow so we can identify them.
I would be open to a config option to set that behavior. I don't actually work with bullmq on any of my own projects anymore, so I don't use this instrumentation myself anymore, and I don't have a good way to exercise it. But I'd be happy to accept a PR.
Hi @JVMartin,
In case you find this useful, we've implemented this change (along with other changes) in our fork of the instrumentation, over at https://github.com/appsignal/opentelemetry-instrumentation-bullmq.