fix(ci): run Langfuse v4 in dual-write mode

[hassiebp]

What changed

• run the Langfuse v4 web and worker containers with LANGFUSE_MIGRATION_V4_WRITE_MODE=dual
• keep the existing SDK implementation and test assertions unchanged

Why

Langfuse platform v4 is now pulled by default in CI, where the default write mode is events_only. The Python SDK test suite still intentionally exercises APIs backed by the legacy trace and observation tables, including the experiment runner's dataset-run-item POST request.

Dual-write mode populates both the v4 events tables and the legacy tables, allowing CI to validate the current SDK contract against the v4 deployment while the remaining platform compatibility issue is fixed separately.

Scope

This is only a CI deployment configuration change. It does not migrate SDK ingestion, batch evaluation, experiment handling, support helpers, or test expectations to v4-native APIs.

Validation

• confirmed the Compose override resolves LANGFUSE_MIGRATION_V4_WRITE_MODE: dual for both langfuse-web and langfuse-worker
• ran an isolated Langfuse v4 stack with the override
• passed 7 representative E2E tests from the original CI failures:
  • trace ingestion and reads
  • observation listing
  • local and server-backed experiment runs
  • multiple experiments on one dataset
  • persisted experiment scores

CI is the authoritative full-suite validation.

Linear

LFE-14622: Run Python SDK CI against Langfuse v4 in dual-write mode

[github-actions]

@claude review

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit:e63e5cbb9f

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".