Cloudflare Workflow reports runtime cancellation after successful run during Sentry teardown #23111

Environment

• @sentry/cloudflare 10.69.0
• Cloudflare Workflows

Problem

The Workflow completes successfully, but the Workers runtime subsequently reports:

The Workers runtime canceled this request because it detected that your Worker code had hung and would never generate a response.

Evidence

For one staging invocation:

• workflow_complete: 2026-08-06T15:02:11.150Z
• finalization_outcome: 2026-08-06T15:02:11.572Z
• workflow_wrapper_outcome with final status SUCCESS: 2026-08-06T15:02:11.587Z
• Workers runtime cancellation: 2026-08-06T15:02:13.572Z

The cancellation occurred approximately 1.985 seconds after the successful wrapper outcome, matching the 2-second Sentry flush timeout. Application work and finalization had completed; the cancellation followed Sentry teardown scheduled through waitUntil.

Relationship to #23005 and #23017

Issue #23005 identified a Workflow-instance reuse deadlock caused by capturing Sentry instrumented waitUntil. PR #23017 fixed that by using getOriginalWaitUntil(context). The post-success cancellation still occurs.

We are testing a local mitigation that awaits flushAndDispose(client) before the Workflow run() returns, instead of scheduling Sentry teardown through waitUntil.

Expected behavior

A successfully completed Workflow should not produce a later Workers runtime cancellation, while Sentry teardown remains safely flushed and disposed.

Please investigate whether Workflow instrumentation should await teardown, use a different lifecycle mechanism, or otherwise prevent Sentry teardown from being treated as hung by the Workers runtime.

Related: #23005, #23017