Hacktoberfest 2026: die Issues, die Maintainer für den Oktober markiert haben – offen und einsteigerfreundlich. Hacktoberfest-Issues durchsuchen

[Bug] opentelemetry-api 1.45 removed `_ExtendedAttributes`, and CI can't resolve 1.45 to notice

Geschlossen
#2,029 1 Kommentar 0 Reaktionen 0 zugewiesene Personen Auf GitHub ansehen

Maintainer antworten meist innerhalb von 1 Tag

@tconley1428 arbeitet bereits daran.

Seit 28.10.2026.

  • #2035 von @tconley1428 — offen

Bewertung

Schwierigkeit
3/5
Geschätzter Aufwand
1-2 Tage
Anfängerfreundlichkeit
56/100
Issue-Typ
Bug
Klarheit
Größtenteils klar
Aktivitätsstatus
Aktiv
Tech-Stack
python

Rechercherichtung

Read temporalio/contrib/opentelemetry/_logger_provider.py and tests/contrib/opentelemetry/test_logger_provider.py, then check how the dependency lock and test-latest-deps resolve OpenTelemetry alongside google-adk. Confirm the typing and test collection failures against opentelemetry-api 1.45. Decide how to cover the newer API despite the shared dependency cap; done when the provider checks and tests work with 1.45 and CI can detect this compatibility gap.

Vom Indexierungsmodell aus dem Issue-Text verfasst.

Beschreibung

bug sdk-python
What are you really trying to do?

Use temporalio.contrib.opentelemetry with opentelemetry-api 1.45, which the opentelemetry extra allows (>=1.26,<2).

Describe the bug

opentelemetry-api 1.45.0 removed the private alias opentelemetry.util.types._ExtendedAttributes. LoggerProvider.get_logger now annotates attributes as Attributes.

  • temporalio/contrib/opentelemetry/_logger_provider.py imports it under TYPE_CHECKING for ReplaySafeLoggerProvider.get_logger. Importing the module still works, but pyright, basedpyright and mypy all report the import on 1.45.
  • tests/contrib/opentelemetry/test_logger_provider.py imports it at runtime, so the module fails to import on 1.45 and none of its tests run.

CI doesn't see this. google-adk>=2.8.0 requires opentelemetry-sdk>=1.39,<=1.42.1, and the project's lock resolves every extra together, so opentelemetry stays at 1.42.1 everywhere. That includes test-latest-deps, which resolved and installed opentelemetry-api==1.42.1 on the October 27 run. test_wrapper_abc_drift.py is meant to catch new ABC methods there, but it can't while that cap holds.

The runtime break from the same release, the new abstract Logger.enabled method, is a separate PR I'm opening alongside this issue.

Minimal Reproduction

On main (64b7005):

$ uv lock --dry-run -P opentelemetry-api==1.45.1 -P opentelemetry-sdk==1.45.1
  × No solution found when resolving dependencies ...
  ╰─▶ Because google-adk>=2.8.0 depends on opentelemetry-sdk>=1.39,<=1.42.1 ...

With only opentelemetry-api installed, 1.44.0 imports the alias and 1.45.1 doesn't:

$ pip install opentelemetry-api==1.45.1
$ python -c "from opentelemetry.util.types import _ExtendedAttributes"
ImportError: cannot import name '_ExtendedAttributes' from 'opentelemetry.util.types'

test_logger_provider.py runs that import at module level, so on 1.45 it fails at collection. I'd expect it to collect and run, as it does with 1.42.1.

Environment/Versions
  • OS and processor: macOS, Apple Silicon
  • Temporal Version: temporalio 1.34.0 and main (64b7005), Python 3.12, opentelemetry-api and opentelemetry-sdk 1.45.1
  • Are you using Docker or Kubernetes or building Temporal from source? Neither
Additional context

Typing attributes as Any in the override and in the test's provider checks against both the older _ExtendedAttributes signature and the 1.45 Attributes one, at the cost of the precise type. While the shared resolution stays capped, an isolated check against the latest opentelemetry could cover the gap the drift test can't see. I can send a PR for the annotation once you pick a direction, or follow a different one if you'd rather keep a precise type.

Vorherrschende Sprache
Python
Sterne
1.2k
Forks
276
Ø Merge
2 T. 2 Std.
Gemergte PRs (30 T.)
87

Entwicklungsumgebung

Erste Schritte

  1. Lesen Sie das ganze Issue und danach den Beitragsleitfaden des Projekts.
  2. Schreiben Sie ins Issue, dass Sie es übernehmen — das erspart doppelte Arbeit.
  3. Forken Sie das Repository und arbeiten Sie in einem Branch.
  4. Öffnen Sie einen Pull Request, der die Issue-Nummer nennt.

Mehr aus temporalio/sdk-python

Alle Issues in temporalio/sdk-python

Ähnliche Issues

Weitere Issues zu Python

Neue Issues direkt in Ihr Postfach

Eine kurze Übersicht über anfängerfreundliche GitHub-Issues.