PrefectHQ/prefect - CodSpeed

prefect

Blog Docs Changelog

Performance History

Latest Results

Update runner apply test for job_variables clearing on upsert (OSS-8269) Co-authored-by: Alexander Streed <alex.s@prefect.io>

devin1/oss-8269-deployment-job_variables-cannot-be-cleared-an-empty-dict-is

39 minutes ago

Add useDeploymentForm tests for clearing job_variables (OSS-8269) Co-authored-by: Alexander Streed <alex.s@prefect.io>

devin1/oss-8269-deployment-job_variables-cannot-be-cleared-an-empty-dict-is

53 minutes ago

Assert scheduled worker runs carry their state (OSS-8268)

devin/1790286145-sqlalchemy-2.1-compat

2 hours ago

Cap SQLAlchemy below 2.1 SQLAlchemy 2.1.0 breaks deployment scheduling ("Can't evaluate bulk DML statement" in _insert_scheduled_flow_runs), and fresh installs on Python 3.11+ resolve to it under the previous <3.0.0 bound. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>

B1aZer:fix/cap-sqlalchemy-2.1

3 hours ago

move url_for import in ReceivedEvent.url to module scope `prefect.utilities.urls` only imports `prefect.events.schemas.events` under TYPE_CHECKING, so a module-level import does not create a cycle. This matches the server-side ReceivedEvent, which already imports `url_for` at module scope.

ethanstoner:fix/client-received-event-url

9 hours ago

fix(runner): stop truncating runner names that contain dots `Runner.__init__` ran every name through `Path(name).stem`, so a runner named `etl.hello` became `etl` and `my-runner-v1.2.3` became `my-runner-v1.2`. Because `Runner.add_flow()` defaults the deployment name to the runner name, the truncation also leaked into deployment names created through `Runner`/`flow.serve()`. Only strip the suffix when a filename was actually passed (`.py`), mirroring `RunnerDeployment.validate_name`, which was fixed the same way in #16553. Fixes #13528

fbossiere:fix/issue-13528-runner-name-dots

14 hours ago

Fix nested schema `required` leaking onto sibling properties in `process_properties` Pass the nested object's required list directly to the recursive call instead of rebinding the loop variable, and forward `allow_none_with_default`. Closes #23189 / OSS-8263 Co-authored-by: Alexander Streed <alex.s@prefect.io>

devin/1790219976-nested-required-leak

20 hours ago

fix: retry immediately instead of crashing when the backoff delay is zero exponential_backoff_with_jitter() handed average_interval straight to clamped_poisson_interval(), which divides by it, so retry_async_fn(..., base_delay=0) raised ZeroDivisionError from inside the exception handler and buried the original error. This mirrors the guard the retry policy already applies in core_policy.

Lesereingrape:fix/retry-backoff-zero-delay

21 hours ago

Latest Branches

0%

Allow clearing deployment `job_variables` with an empty dict#23203

50 minutes ago

84ea1d2

devin1/oss-8269-deployment-job_variables-cannot-be-cleared-an-empty-dict-is

0%

Make the server compatible with SQLAlchemy 2.1#23201

2 hours ago

d65d82a

devin/1790286145-sqlalchemy-2.1-compat

0%

Cap SQLAlchemy below 2.1#23200

3 hours ago

8851e7b

B1aZer:fix/cap-sqlalchemy-2.1

© 2026 CodSpeed Technology

Home Terms Privacy Docs