Does anyone have experience with DataDog and BentoML integration? Looking through the Slack history, I see a few posts about it, but nothing with super concrete details. So far, I’ve only naively tried starting a BentoML model service with

ddtrace-run bentoml serve

, which doesn’t seem to work (this was with

bentoml==1.1.11

). Curious if anyone has any suggestions. Will keep troubleshooting on my own as well.

Have you tried using dd’s TraceMiddleware with BentoML’s

svc.add_asgi_middleware

API?

from ddtrace.contrib.asgi import TraceMiddleware

Thanks for the suggestion. I was on vacation, but will try this out. Do you know if this would instrument both the service and the runners? Is that distinction still relevant with

1.2.x

versions? I’ve been exploring the

1.1.x

series and haven’t yet tried out

1.2.x

, which seems pretty significantly different.