/v1/videos dispatches call_type avideo_generation, which CallTypes did not
know and no guardrail translation handler covered, so the unified guardrail
hook returned the request unscanned. Add the video call types and an OpenAI
video guardrail translation package that scans the prompt for create, remix,
edit and extension requests
Resolves LIT-6685
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
Two CI checks were asserting behaviour the proxy no longer has. Neither
was catching anything; both now fail for the right reason.
budget_exceeded (tests/otel_tests/test_e2e_budgeting.py)
bf804f5188 made 422 the default for budget refusals and added
budget_exceeded_status_code to restore 429 for callers that need it. The
e2e budget tests still asserted 429, so all six have been failing on a
status change that was deliberate. Assert 422, the documented default,
rather than reading litellm.budget_exceeded_status_code back — a test
that asks the code what it does would have passed straight through this
change and through the next one.
The helpers also caught bare Exception, so a connection error reached
`e.body` and surfaced as an AttributeError instead of a failed
assertion. Narrow both to openai.APIStatusError, which is what a refusal
actually raises (UnprocessableEntityError for 422, RateLimitError for
429), and let anything else propagate as itself.
test_bad_database_url (.circleci/config.yml)
The check required "Database setup failed after multiple retries", which
only the v1 resolver emits, OR uvicorn's "Application startup failed.
Exiting.". With v2 the default, the first branch is dead and the whole
assertion rests on an incidental uvicorn line that HEAD's run did not
emit at all.
Assert the behaviour instead of the wording: the container exits
non-zero, the log names the unreachable server (P1001), it never reaches
"Application startup complete", and it is not left running. The exit
code was previously discarded by `|| true`, so the one thing the job
most needed to check was never checked.
Verified by running the bad-DATABASE_URL container: exit 3, P1001
present, no startup-complete line, container stopped — the new check
passes and the old one passed only by the uvicorn line's accident.
Register s3_access_key_id, s3_secret_access_key and s3_encryption_key_id as
LiteLLM-owned batch params so they are no longer forwarded to Bedrock as
additionalModelRequestFields (which 400s ordinary chat on a batch-configured
deployment), keep them on CredentialLiteLLMParams so the batch/file paths
still receive them, and redact the S3 credential key names in debug logs.
Resolves LIT-8290
Co-authored-by: yucheng <yucheng@berri.ai>
Co-authored-by: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
The router coverage gate wants every router.py function reached by name from
a router test. The two per-endpoint attempt functions were only reached through
their callers, so each now has a direct test proving the per-request controls
carrier never reaches the provider call and every hop's stream comes back wrapped.
Combines the Redis cluster bridge changes with the Azure Blob variant and
keeps account URL query parameters when building the container URL
Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
The /v1/responses and /v1/messages streaming wrappers only ever wrapped the
primary's stream, so a hop reached through the regular fallback chain had no
mid-stream handler: its failure re-raised, or the outer wrapper retried the
same entry with a fresh attempted set and never reached the rest of the list.
Every attempt of the chain now runs through a per-endpoint attempt function
that wraps its own stream, mirroring chat completions, and the per-request
fallback and retry overrides ride a frozen carrier so each hop's re-entry
still sees them after the retry layer pops them.
