[opentelemetry-instrumentation-genai-portkey] Add initial boilerplate / setup

[DylanRussell]

Description

This PR adds initial boilerplate / setup stuff. This is another open inference library port..

Portkey wraps around the model APIs (instead of the SDKs which we instrument) and provide the user with a single interface to call any model provider API. The portkey SDK sends the request to portkey's endpoint where they do the translation to the API call (adds the users API key) and does various other stuff.. It then makes the call and returns the response to the user.. They support streaming and sync/async. There are just 4 methods which we need to monkey patch on the portkey SDK..

This instrumentation will just emit inference spans..

Type of change

• New feature (non-breaking change which adds functionality)

How has this been tested?

Unit tests

Checklist

• Followed the style guidelines of this project
• Changelog updated if the change requires an entry
• Unit tests added
• Documentation updated

[Copilot]

Pull request overview

This PR introduces a new instrumentation package, opentelemetry-instrumentation-genai-portkey, and wires it into the repo’s workspace, tox matrix, and typecheck scope. The package scaffolding includes an instrumentor, patching entrypoints, tests, conformance harness stubs, documentation, and examples, but the actual telemetry mapping is currently unimplemented.

Changes:

• Add the new opentelemetry-instrumentation-genai-portkey package (project metadata, instrumentor, patch/unpatch plumbing, README, license, changelog fragment).
• Wire Portkey into the monorepo (uv workspace lock entry, root tox envs, root pyright include/excludes).
• Add initial tests, conformance test harness, and manual examples.

Reviewed changes

Copilot reviewed 24 out of 25 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
uv.lock	Adds the new Portkey instrumentation package to the uv workspace lock and resolves new transitive deps.
tox.ini	Adds {oldest,latest,conformance} test envs for Portkey and installs the package in typecheck deps.
pyproject.toml	Includes Portkey in pyright include and excludes its tests/examples from typechecking.
instrumentation/opentelemetry-instrumentation-genai-portkey/pyproject.toml	Defines the new package’s metadata, deps, extras, and entrypoint registration.
instrumentation/opentelemetry-instrumentation-genai-portkey/src/opentelemetry/instrumentation/genai/portkey/init.py	Adds PortkeyInstrumentor and wires TelemetryHandler + patching.
instrumentation/opentelemetry-instrumentation-genai-portkey/src/opentelemetry/instrumentation/genai/portkey/patch.py	Adds Portkey SDK monkey-patch hooks (currently placeholder wrappers).
instrumentation/opentelemetry-instrumentation-genai-portkey/src/opentelemetry/instrumentation/genai/portkey/package.py	Declares the instrumented dependency string for instrumentation_dependencies().
instrumentation/opentelemetry-instrumentation-genai-portkey/src/opentelemetry/instrumentation/genai/portkey/version.py	Adds package version module.
instrumentation/opentelemetry-instrumentation-genai-portkey/README.rst	Documents installation/usage and completion-hook configuration.
instrumentation/opentelemetry-instrumentation-genai-portkey/LICENSE	Adds Apache-2.0 license file for the new package.
instrumentation/opentelemetry-instrumentation-genai-portkey/CHANGELOG.md	Adds towncrier-managed changelog scaffold for the package.
instrumentation/opentelemetry-instrumentation-genai-portkey/.changelog/443.added	Adds a changelog fragment announcing Portkey instrumentation.
instrumentation/opentelemetry-instrumentation-genai-portkey/tests/conftest.py	Registers shared test-util-genai fixtures + VCR config and provides instrument/uninstrument fixtures.
instrumentation/opentelemetry-instrumentation-genai-portkey/tests/test_instrumentor.py	Adds basic unit tests for instrumentor construction and lifecycle.
instrumentation/opentelemetry-instrumentation-genai-portkey/tests/test_completion_hook.py	Tests that completion_hook is forwarded and defaults to load_completion_hook().
instrumentation/opentelemetry-instrumentation-genai-portkey/tests/test_conformance.py	Adds conformance test entrypoint using run_conformance(...).
instrumentation/opentelemetry-instrumentation-genai-portkey/tests/requirements.oldest.txt	Adds oldest-factor requirements file (currently only explanatory comments).
instrumentation/opentelemetry-instrumentation-genai-portkey/tests/requirements.latest.txt	Adds latest-factor requirements file and editable installs for local packages.
instrumentation/opentelemetry-instrumentation-genai-portkey/tests/init.py	Marks tests as a package.
instrumentation/opentelemetry-instrumentation-genai-portkey/tests/conformance/init.py	Marks conformance tests as a package.
instrumentation/opentelemetry-instrumentation-genai-portkey/tests/conformance/chat.py	Adds a chat conformance scenario using VCR cassette replay.
instrumentation/opentelemetry-instrumentation-genai-portkey/examples/manual/requirements.txt	Adds dependencies for the manual example.
instrumentation/opentelemetry-instrumentation-genai-portkey/examples/manual/README.rst	Documents how to run the manual examples.
instrumentation/opentelemetry-instrumentation-genai-portkey/examples/manual/main.py	Adds a manual instrumentation example that performs a chat completion.
instrumentation/opentelemetry-instrumentation-genai-portkey/examples/manual/custom_hook.py	Adds a manual example showing a custom CompletionHook.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[opentelemetry-pr-dashboard]

Pull request dashboard status

Waiting on the author · refreshed 2026-08-22 18:37 UTC

Respond to 2 review items (e.g. link a commit, explain why not, ask a follow-up):

• Top-level threads: 1, 2

Status above doesn't look right?

• Just replied or pushed? Anything around or after the refresh time above may not be picked up yet — give it a few minutes.
• Should this be with reviewers? Comment /dashboard route:reviewers to route it to them.
• Anything wrong — including the routing? Report it with what you expected; it helps us improve the dashboard.

[lmolkova]

@DylanRussell do you want to try stacked PRs ? They don't work on forks, but you should be able to push to origin with your github alias as prefix and then they start working.

I'm a bit worried that we created a lot of empty skeletons, but we'renot populating them with instrumentations and I'd prefer us to merge real instrumentations. Stacked PRs allow to review individual stages in isolation, but merge them together

[eternalcuriouslearner]

Question(non-blocking): Should we emit inference spans? Can we just emit agent, workflow and tool spans?