Manually entered Pushover aliases were normalised on save but not on test, so the test could exercise a different payload. Apply the same normalisation and retain a failing-before parity regression; 56 focused webhook tests pass.
Change-source: pulse-maintainer
pulse-triage[bot]