ping-javascript-sdk
Failed after Nnx-cloud[bot]
View latest

fix(protect): update-signals-sdk [Self-Healing CI Rerun]

863d7619b2Open on GitHub

Self-Healing CI

Classification: Flaky task

We looked at the failing WebAuthn e2e test (@forgerock/journey-suites) and found it unrelated to our changes. Our PR only updates the Protect signals SDK bundle, a changeset, and regenerated api-reports, and the failure is a 5s timeout waiting for the Logout button after the authenticate step. The task has a non-zero flakiness rate and no matching failure on other branches, so we recommend a rerun with no code changes.

An empty commit was automatically applied to the branch to trigger a new CI pipeline execution to resolve the flaky task.

This failure was classified as flaky_task, not code_change, for these reasons:

  1. The PR does not touch the failing code path.
  • The PR updates the Protect signals SDK bundle (packages/protect/src/lib/signals-sdk.js).
  • It also adds a changeset and regenerates api-report files.
  • None of these affect the WebAuthn journey flow or the journey-suites e2e app.
  1. The error does not reference changed code.
  • The failing test is the journey-suites WebAuthn register/authenticate/delete test.
  • The Registration step passed.
  • The 'Authenticate with the registered WebAuthn device' step then timed out after 5000ms waiting for the Logout button to be visible.
  • No changed function, import or variable appears in the error.
  1. The failing project is not touched.
  • @forgerock/journey-suites is not in touched_projects.
  1. The failure looks timing-related.
  • It is a UI wait against a Chromium virtual authenticator, which can be timing-dependent.
  • The task has a non-zero flakiness rate (about 1.4%).
  1. The failure does not appear on other branches.
  • The similar-task-failure-detector subagent returned no matching failures elsewhere.
  • That rules out a pre-existing failure on the base branch, so environment_state does not apply.
  • It also means the failure has not been seen elsewhere, so the evidence for flakiness is limited to the low non-zero flakiness rate and the unrelated diff.

No code changes are proposed. A rerun is the appropriate action.