Wrapture monkey-patches Python for both tests and tracing

Graham Dumpleton released Wrapture, a Python library that monkey-patches methods, attributes, dicts, and generators for two jobs at once. It replaces separate toolchains, mocking libraries for tests and APM agents for production, with a single patch layer.

In tests, it records call sequences as timelines you can assert against, the job a mocking framework usually does. Point it at a live application through a TOML config instead, with no changes to the source. It traces the same code path and exports to OpenTelemetry. Flask, Django, FastAPI, and SQLAlchemy get instrumentation built in.

Simon Willison wrote up the release in a post on Wrapture. He calls it "a new monkey patching library that serves both testing and observability (think New Relic style tracing) at the same time." He also flags the split status. Wrapture is alpha, though Dumpleton calls it production-ready.

That gap between alpha and production-ready matters because tests and production tracing fail in different ways. A broken assertion in CI is cheap to fix. A patched method misbehaving under real traffic is not. Anyone plugging OpenTelemetry export into a live Flask or Django app is running ahead of the library's own confidence.