Use experiment's metrics API

Author: owbone

tests/integrations/test_standalone_lifecycle.py

@@ -257,8 +257,6 @@ def test_resume_experiment():
 @pytest.mark.cloud()
 def test_resume_old_experiment():
     """Resume an experiment that has many newer experiments in the same teamspace."""
-    from litlogger.api.metrics_api import MetricsApi
-
     target_name = f"standalone_resume_old-{uuid.uuid4().hex}"
     filler_prefix = f"standalone_resume_old_filler-{uuid.uuid4().hex}"
 
@@ -269,10 +267,8 @@ def test_resume_old_experiment():
     project_id = exp1._teamspace.id
     target_stream_id = exp1._metrics_store.id
 
-    # Use an unlimited-retry client so that 429s under parallel test load are
-    # absorbed by the client's exponential backoff instead of failing the test.
     client = LitRestClient(max_retries=-1)
-    api = MetricsApi(client=client)
+    api = exp1._metrics_api
 
     filler_ids: list[str] = []
     try: