Drive real traffic through the proxy and read the events back out of the
Datadog Logs Search API to prove delivery instead of trusting the proxy. The
DatadogClient reads DD_API_KEY, DD_SITE and DD_APP_KEY from the environment and
polls the search API for a per-run marker stamped into each request
Three scenarios cover the contract: a batch of gpt-5.5 chat completions shipped
with no drops, the /v1/responses path on a claude-haiku-4-5 deployment
exercising the Responses-to-Anthropic translation, and a failed request that
must land as an error event. The e2e compose config enables the datadog (and
prometheus) callback so the delivery path is actually exercised, and the suite
skips unless all three credentials are set
* chore(e2e): untrack gateway config and document e2e test location
Stop tracking tests/e2e/gateway/litellm-config.yml so the local proxy config stays on the machine
Add a note to CLAUDE.md that new e2e tests belong in tests/e2e/ and must follow that directory's conventions
* chore(e2e): add self-contained docker compose stack for local runs
Ship a docker-compose.yml that starts the proxy with a throwaway Postgres and Redis and inlines the proxy config with example models, so contributors can bring up a local gateway with nothing but a .env. Update CONTRIBUTING.md to match the inline-config flow
* chore(e2e): drop the second gemini deployment; one key is enough locally
* docs(e2e): make pre-commit steps ordered and require flagging internally found issues
