mirror of
https://github.com/BerriAI/litellm.git
synced 2026-10-12 03:48:56 +00:00
79209b92a1
* test(e2e): add enum values, auto-discovering label gates and secret hiding for e2e metadata * test(e2e): tag guardrails and logging tests with Subject metadata and record client steps * test(e2e): leave the guardrails and logging harness unit tests untagged * test(e2e): let the inner create_model step name the guardrail backend deployment * docs(e2e): name every markerless harness test file that carries no Subject * test(e2e): keep the step discovery comprehensions to one for clause * test(e2e): declare the default guardrail backend model on the tests that drive it
68 lines
2.3 KiB
Python
68 lines
2.3 KiB
Python
from __future__ import annotations
|
|
|
|
import time
|
|
|
|
import pytest
|
|
from prometheus_client.parser import text_string_to_metric_families
|
|
|
|
from e2e_config import unique_marker
|
|
from e2e_metadata import Domain, Mode, Provider, Route, Subject, meta
|
|
from lifecycle import ResourceManager
|
|
from logging_client import LoggingClient
|
|
|
|
pytestmark = pytest.mark.e2e
|
|
|
|
DRIVER_MODEL = "gemini-2.5-flash"
|
|
QUEUE_TIME_METRIC = "litellm_request_queue_time_seconds"
|
|
ALIAS_LABEL = "api_key_alias"
|
|
|
|
|
|
def _observation_count(exposition: str, alias: str) -> float | None:
|
|
return next(
|
|
(
|
|
sample.value
|
|
for family in text_string_to_metric_families(exposition)
|
|
for sample in family.samples
|
|
if sample.name == f"{QUEUE_TIME_METRIC}_count" and sample.labels.get(ALIAS_LABEL) == alias
|
|
),
|
|
None,
|
|
)
|
|
|
|
|
|
class TestPrometheusRequestQueueTime:
|
|
@pytest.mark.covers("logging.prometheus.success.records_queue_time")
|
|
@meta(
|
|
Subject(
|
|
domain=Domain.OBSERVABILITY,
|
|
route=Route.METRICS,
|
|
providers=(Provider.GEMINI,),
|
|
models=(DRIVER_MODEL,),
|
|
mode=Mode.NONSTREAM,
|
|
)
|
|
)
|
|
def test_queue_time_histogram_records_an_observation(
|
|
self, client: LoggingClient, resources: ResourceManager
|
|
) -> None:
|
|
alias = f"e2e-queue-time-{unique_marker()}"
|
|
key = client.key_with_alias(alias, models=[DRIVER_MODEL])
|
|
resources.defer(lambda: client.delete_key(key))
|
|
|
|
response = client.chat(key, DRIVER_MODEL, f"reply with one word {alias}")
|
|
assert response.model, f"driver call returned no model: {response}"
|
|
|
|
deadline = time.monotonic() + client.proxy.poll_timeout
|
|
count: float | None = None
|
|
while time.monotonic() < deadline:
|
|
count = _observation_count(client.scrape_metrics(), alias)
|
|
if count is not None and count > 0:
|
|
break
|
|
time.sleep(client.proxy.poll_interval)
|
|
|
|
assert count is not None, (
|
|
f"{QUEUE_TIME_METRIC} has no series for {ALIAS_LABEL}={alias}; the histogram was "
|
|
f"never observed for a request that succeeded"
|
|
)
|
|
assert count > 0, (
|
|
f"{QUEUE_TIME_METRIC} series for {alias} exists but recorded {count} observations; "
|
|
f"the metric is registered yet never written"
|
|
)
|