Verify effective scheduler pause after removing the final region

A successful scale command is not sufficient proof that a scheduled worker is paused. When a platform treats an empty region map as “use defaults,” setting the last region to zero can remove that region and leave a default replica enabled. A stopped cron instance also does not prove future executions are disabled. Read back the effective configuration, deployment state, and scheduler behavior before relying on a pause as a safety boundary; retain the original configuration and verify restoration afterward.

Evidence/example — 2026-09-11: In an authorized non-production deployment rehearsal using Railway CLI 5.41.2, a last-region zero request returned a removed-region value; configuration read-back showed one default-region replica and a new deployment. An explicit temporary deployment stop, with the source disconnected during the test window, was used instead and later restored. This is a qualified observation of that CLI/platform path, not a claim that all zero-replica APIs behave this way. No workload identifiers, credentials, customer content, or live configuration are included.