mirror of
https://github.com/BerriAI/litellm.git
synced 2026-11-05 07:37:50 +00:00
b18188dee2
* test(e2e): retire the quiet_stack lock quiet_stack held an exclusive host-wide lock for six tests, draining every other xdist worker while they ran. On the full e2e lane that left 35-38% of worker time idle. Each of those tests now lives where it can run without pausing the suite - the two memory tests move to perf/ behind the perf marker, so the perf lane runs them on its own fresh stack. Their refusing deployments use a literal key because the perf stack carries no provider keys - the stored-request size check from the memory test becomes an integration test on an owned proxy that stores prompts - the discount and margin spend-row tests are dropped from e2e. Their breakdown assertions join the existing integration pricing test, which already owns the discount and margin repricing flow - the Bedrock guardrail cost test stays in e2e without the pricing reset fixture, since nothing else on the stack writes pricing config now - the MCP semantic tool filter test turns on a proxy-wide setting, so it gets a new proxy_wide_config opt-in marker and runs after the main run * test(e2e): register the claude_code deployments as one batch The claude_code session fixture registered its 24 deployments one at a time on every xdist worker, and each one waited out the 15 s propagation window on its own, about 6 minutes per worker. ProxyClient.register_models posts the whole batch, polls each model until every replica serves it, then waits out propagation once from the last write. A model that fails to post or never becomes servable comes back as ModelNotRegistered, and an unservable one is deleted, so the fixture keeps reporting per-deployment failures the way it did
23 lines
3 KiB
INI
23 lines
3 KiB
INI
[pytest]
|
|
# Config when any e2e suite under tests/e2e/ is run directly, e.g.
|
|
# uv run pytest tests/e2e/quota_management/spend_tracking/ -v
|
|
# The e2e marker is also registered in conftest.py for runs rooted elsewhere.
|
|
addopts = --strict-markers --strict-config --reruns 1 --only-rerun "kind='network'" --only-rerun "status_code=5[0-9][0-9]"
|
|
markers =
|
|
e2e: live test that requires a running proxy and real provider keys
|
|
meta: typed e2e_metadata.Subject describing what this test drives (domain/route/providers/models/capabilities/mode); attach it with @meta(Subject(...)), never as a bare pytest.mark
|
|
replayable: edge-wired test whose provider traffic replays from a fixture bundle, so it makes zero provider calls in replay mode; the record/replay CI lane selects it with -m replayable
|
|
perf: performance test (throughput, latency, memory); collected last so it never perturbs latency-sensitive suites, and deselected unless E2E_PERF_TESTS is set
|
|
weekly: real-provider anomaly load test that spends real money; deselected unless E2E_WEEKLY_ANOMALY is set
|
|
managed_files: needs a proxy running with require_managed_files enabled; deselected unless E2E_MANAGED_FILES_STACK is set
|
|
prompt_caching_stack: needs a proxy running with router_settings.optional_pre_call_checks including prompt_caching; deselected unless E2E_PROMPT_CACHING_STACK is set
|
|
cli_determinism: drives the real claude CLI for several seconds; deselected unless E2E_CLI_DETERMINISM is set
|
|
redis_chaos: load test that pauses the proxy's Redis outright mid-run; needs a proxy booted from gateway/redis_chaos_ci_config.yml on the same host, and is deselected unless E2E_PERF_TESTS is set
|
|
provider_edge_host: routes provider traffic through the pytest host's edge in every fixture mode, so the gateway must reach the pytest host; deselected unless E2E_PROVIDER_EDGE_HOST_REACHABLE is set
|
|
owned_gateway: boots its own proxy from source against the stack's Postgres, so it needs DATABASE_URL on the pytest host; deselected unless E2E_OWNED_GATEWAY is set
|
|
otel_v2: needs a proxy running with LITELLM_OTEL_V2=true; deselected unless E2E_OTEL_V2 is set
|
|
otel_tls: needs a stack whose gateway exports OTLP over TLS signed by the CA in SSL_CERT_FILE; deselected unless E2E_OTEL_EXPORTER_ENDPOINT is set
|
|
secret_manager: needs a proxy booted from gateway/secret_manager_<system>_ci_config.yml against that live secret manager; deselected unless E2E_SECRET_MANAGER names the backend (see secret_manager/secret_backends.py)
|
|
sagemaker_nova: needs a deployed SageMaker Nova endpoint the proxy's AWS credentials can invoke; deselected unless SAGEMAKER_NOVA_ENDPOINT names it
|
|
sagemaker_nova2_lite: needs a deployed SageMaker Nova 2 Lite endpoint the proxy's AWS credentials can invoke; deselected unless SAGEMAKER_NOVA2_LITE_ENDPOINT names it
|
|
proxy_wide_config: turns on a proxy-wide setting through /config/update, so it runs on a stack no other suite is driving; deselected unless E2E_PROXY_WIDE_CONFIG_STACK is set
|