Set up consecutive failure alerting

Query for the previous run's consecutive_failure_count using: `SELECT consecutive_failure_count FROM cron_execution_logs ORDER BY started_at DESC LIMIT 1 OFFSET 1`. Increment by 1 on failure, set to 0 on success. Only dispatch the alert when the new count equals exactly 2 (the threshold). Use Supabase Vault `vault.decrypted_secrets` to retrieve the webhook URL at runtime.

Implement the alert dispatch as a separate async function called with `await` but wrapped in its own try/catch so alert delivery failure does not pollute the cron execution status. Structure the Slack message using Block Kit for readability: include a header, error details section, and a 'Last successful run' context block. If using email, use a transactional email provider already approved by the project — confirm with the team before integrating a new service.

Unit tests: mock the alerting channel call and assert it is invoked exactly once when consecutive_failure_count transitions to 2, and NOT invoked for count 1 or count 3+. Assert alert payload structure matches specification (error_message, last_success_at, consecutive_failure_count). Integration tests: simulate two consecutive failed cron runs in a test Supabase environment and verify the alert channel mock received exactly one call with correct payload. Test the reset path: after two failures followed by a success, assert consecutive_failure_count is 0 in the new log row.

Test the timeout path: simulate a slow webhook endpoint and verify the cron function still completes within acceptable time.