scriptoriumadmin/litellm
1
0
Fork 0
mirror of https://github.com/BerriAI/litellm.git synced 2026-09-26 01:12:21 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 145

test(ci): mark Railway-dependent router/overhead tests as flaky

Browse source 
Two more CircleCI tests intermittently fail with the Railway-hosted
fake openai endpoint (https://exampleopenaiendpoint-production.up.railway.app)
returning 404 'Application not found' on cold-start:

  - tests/local_testing/test_router.py::test_router_text_completion_client
    (litellm_router_testing job)
  - tests/litellm_utils_tests/test_litellm_overhead.py::test_litellm_overhead_non_streaming[openai/self_hosted]
  - tests/litellm_utils_tests/test_litellm_overhead.py::test_litellm_overhead_stream[openai/self_hosted]
    (litellm_utils_testing job)

Same upstream-flake root cause as PR #28315; mirror that fix by adding
@pytest.mark.flaky(reruns=2, reruns_delay=5) so transient endpoint
unavailability doesn't fail the build.

Co-authored-by: Krrish Dholakia <krrish-berri-2@users.noreply.github.com>
This commit is contained in:
Cursor Agent 2026-05-20 10:11:26 +00:00
parent e59e34bed3
commit dbcc67a603
No known key found for this signature in database
2 changed files with 3 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

2
tests/litellm_utils_tests/test_litellm_overhead.py
View file

@ -75,6 +75,7 @@ async def _vertex_ai_mocks():
yield
@pytest.mark.flaky(reruns=2, reruns_delay=5)
@pytest.mark.asyncio
@pytest.mark.parametrize(
"model",
@ -141,6 +142,7 @@ async def test_litellm_overhead_non_streaming(model):
pass
@pytest.mark.flaky(reruns=2, reruns_delay=5)
@pytest.mark.asyncio
@pytest.mark.parametrize(
"model",

1
tests/local_testing/test_router.py
View file

@ -1628,6 +1628,7 @@ def test_router_add_deployment():
assert new_model_id_list[1] != new_model_id_list[0]
@pytest.mark.flaky(reruns=2, reruns_delay=5)
@pytest.mark.asyncio
async def test_router_text_completion_client():
# This tests if we re-use the Async OpenAI client