Yun Kim
Yun Kim
## Description Currently, users have to create their own Propagator classes and/or use the `HTTPPropagator` along with extra steps of extracting the context from request headers and activating them with...
### Description: I tried installing gevent latest (21.12.0) through pip with Python 3.11, through alpine docker image and encountered a build error, while Python versions 3.10 + 3.9 + 3.7...
I have a celery task that raises an exception: ```python @self.app.task def fn_exception(): raise Exception("Task class is failing") ``` When the task is applied, instead of being caught and sending...
This issue is to track our progress in supporting Python 3.11. The dd-trace-py library currently does not support Python 3.11, which is scheduled to be released October 24 2022. We...
## Description #4302 added a `callonce()` decorator as a cache utility function. However, this imported `inspect.getargspec()` was removed as of Python 3.11. This was fixed by importing `inspect.getfullargspec()`, which is...
## Description Python 3.11 has made some new changes to their C internal API that we need to address in a couple of places, mostly in our profiler. I've created...
Previously DD_LLMOBS_APP_NAME was a strongly recommended (but not required) config for LLMObs users. However, in the case users do not specify the env var, the ML Obs product will not...
Hi there, running into the same issue as #832. It seems like getting [rid of the manual step](https://github.com/kevin1024/vcrpy/commit/5fa7010712fcd8d61a3145545ec2757b1d2d4b25#diff-bd5fad4e275c5a3ee11e3b297c45511db9b9fae64e9cf9c157db250e68d54128) to set `response._content`, as well as mocking the `httpx.Response.read()` function resulted in...
This PR introduces automatic input/output annotation for task/tool/workflow/agent/retrieval spans created using function decorators. Specifically, the input args/kwargs provided to a traced function will be captured as a dictionary of key-value...
This PR changes the names of openai spans submitted to LLMObs to follow `openai.` instead of a very generic `openai.request`, where `OPERATION_ID` is one of `createCompletion, createChatCompletion, createEmbedding`. The goal...