Skip to content

ref(langchain): Remove set_data_normalized for primitive attributes #5510

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
alexander-alderman-webb wants to merge 1 commit into
base: master
Choose a base branch
from
+1 −1
Draft

ref(langchain): Remove set_data_normalized for primitive attributes #5510

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion sentry_sdk/integrations/langchain.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -389,7 +389,7 @@ def on_llm_start(

for key, attribute in DATA_FIELDS.items():
if key in all_params and all_params[key] is not None:
set_data_normalized(span, attribute, all_params[key], unpack=False)
span.set_data(attribute, all_params[key])
Copy link

@cursor cursor bot Feb 23, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Unnormalized tool call data on spans

Medium Severity

Replacing set_data_normalized with span.set_data for all DATA_FIELDS can store non-primitive values (notably function_call/tool_calls) without JSON normalization. This can yield inconsistent encoding versus other paths that still use set_data_normalized, and may break span serialization or downstream consumers expecting a JSON string for SPANDATA.GEN_AI_RESPONSE_TOOL_CALLS.

Fix in Cursor Fix in Web

Copy link

@sentry sentry bot Feb 23, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Bug: The change in on_llm_start removes necessary serialization for complex objects like function_call and tool_calls by replacing set_data_normalized() with span.set_data().
Severity: MEDIUM

Suggested Fix

The logic should be updated to selectively apply span.set_data() only to primitive types within DATA_FIELDS. For complex fields like function_call and tool_calls, continue to use set_data_normalized() to ensure proper serialization.

Prompt for AI Agent 
Review the code at the location below. A potential bug has been identified by an AI
agent.
Verify if this is a real issue. If it is, propose a fix; if not, explain why it's not
valid.

Location: sentry_sdk/integrations/langchain.py#L392

Potential issue: The change in `on_llm_start` replaces `set_data_normalized()` with
`span.set_data()` for all `DATA_FIELDS`. While the intent was to do this for primitive
types, this change also affects complex objects like `function_call` and `tool_calls`.
These fields require serialization to a JSON string, which `set_data_normalized()`
performs but `span.set_data()` does not. Passing these complex objects directly can lead
to serialization failures when the span is transmitted, resulting in lost telemetry
data. This creates an inconsistency with `on_chat_model_start`, which still correctly
uses normalization for the same fields.

Did we get this right? 👍 / 👎 to inform future reviews.


_set_tools_on_span(span, all_params.get("tools"))

Expand Down
Loading