fix(tracing): stop fabricating EndTime for live spans#1806
Merged
Conversation
otel_span_to_uipath_span stamped EndTime=now() when converting a span that has not ended (otel_span.end_time is None), and the UiPathSpan dataclass defaulted end_time to now() as well. This makes in-progress upserts (the upsert_span RUNNING -> OK lifecycle) indistinguishable from ended spans downstream: the LLMOps traceview UI renders phantom sub-millisecond durations for open snapshots, and the Insights OTLP export's unclosed-span filter never fires — so every v3-ingested span reaches customer OTLP endpoints twice (open + terminal snapshot). Verified against LLMOps alpha storage: open writes arrive with EndTime == UpdatedAt ≈ StartTime. Live spans now convert with end_time=None, serialized as EndTime: null — the LLMOps v3 ingest contract (SpanV3Req.EndTime) is nullable, and OTel itself models unfinished spans with end_time=None; the fallback was a local invention, not a framework requirement. Ended spans are unchanged. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
github-actions
Bot
added
test:uipath-langchain
Contributor
There was a problem hiding this comment.
Pull request overview
This PR fixes span conversion/serialization so in-progress (live) OpenTelemetry spans are not treated as ended spans by downstream UiPath tracing consumers. It preserves end_time=None for live spans so "EndTime": null is emitted, avoiding fabricated durations and preventing duplicate exports in pipelines that rely on EndTime == null to detect unclosed spans.
Changes:
- Update
UiPathSpan.end_timeto beOptional[str] = Noneso “not ended” is representable and serializes as JSONnull. - Update
otel_span_to_uipath_spanto preserveend_time=Nonewhen the OTEL span has not ended (instead of stampingdatetime.now()). - Add tests covering live-span (
EndTime is None) and ended-span (real end time preserved) conversion behavior.
Reviewed changes
Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.
| File | Description |
|---|---|
packages/uipath-platform/src/uipath/platform/common/_span_utils.py |
Stops fabricating end times for live spans and makes UiPathSpan.end_time nullable so EndTime: null can be emitted. |
packages/uipath-platform/tests/services/test_span_utils.py |
Adds regression tests ensuring live spans convert with end_time=None and ended spans retain their actual end time. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
danflorescu96
approved these changes
Jul 9, 2026
yashwagle1
approved these changes
Jul 9, 2026
Required by check-versions since 0.2.5 is already published on PyPI. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
🚨 Heads up:
|
|
mjnovice
approved these changes
Jul 9, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
otel_span_to_uipath_spanstampsEndTime = datetime.now()when converting a span that hasn't ended (otel_span.end_time is None), and theUiPathSpandataclass defaultsend_timetonow()too. A live OTel span legitimately hasend_time = None— the fallback is a local invention, not a framework requirement.Why it matters: the fabricated EndTime makes in-progress upserts (the
upsert_spanRUNNING → OKlifecycle) indistinguishable from ended spans everywhere downstream:EndTime == nullspans) never fires, so every v3-ingested span reaches customer OTLP endpoints twice — once as an open snapshot with a ~0–1ms duration and no status, once as the real terminal span.Verified against LLMOps alpha Table Storage: open writes arrive with
EndTime == UpdatedAt ≈ StartTime(e.g. an agentRun open write withStart .456 / End .457 / Status Unsetfollowed 33s later by the terminalOkwrite).Change
end_time = None, serialized as"EndTime": null. The LLMOps v3 ingest contract accepts null (SpanV3Req.EndTimeisDateTime?), and storage/UI already handle unclosed spans on the v2 path.UiPathSpan.end_timebecomesOptional[str] = None(the eval runtime already passesdata.get("end_time")— i.e. possiblyNone— into this field today).Not addressed here (follow-up worth considering)
Some open-write call sites don't pass
status_override=SpanStatus.RUNNING, so open snapshots land withUnsetstatus (observed on Agents-source spans) instead ofRunning. Marking all open upsertsRUNNINGwould make in-progress state explicit rather than inferred. Insights-side interim handling: UiPath/Insights-monitoring#3041.Tests
packages/uipath-platform: full suite 1434 passed, 7 skipped (pre-existing credential-gated skips). Two new tests pin the contract: live span →end_time is None/to_dict()["EndTime"] is None; ended span keeps its real end time.🤖 Generated with Claude Code