Replies: 1 comment
-
I have managed to add my custom exporters using explicit code with OpenTelemetry SDK, but a better way would be to support using custom extensions in Promptflow. Currently, extensions factory is used, but it is hardcoded to only support existing AzureML or Default extensions. A more flexible design would be to allow factory resolve the extension type from config/env var, which will enable implementing custom extensions and have control of how FlowMonitor, logger, etc are created without forking and modifying the Promptflow |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
E.g. I already have a DataDog that supports or want to use existing Zipkin in my Dev environment.
Currently, it seems there's no way around it as there's a hardcoded dependency that forces using either AzureML or local log as trace destination, so I cannot leverage the fact that PromptFlow is instrumented with OpenTelemetry, and I can't use existing services that support it too.
Beta Was this translation helpful? Give feedback.
All reactions