Background job failures are logged and retried

Background jobs that fail silently — especially those processing payments, sending billing notifications, or syncing shared data — can fail for hours before anyone notices. CWE-391 (Unchecked Error Condition) applies directly: a Vercel Cron route that throws without a try/catch produces no log and no retry path. CWE-778 (Insufficient Logging) applies when errors are caught but not recorded. ISO 25010 reliability.fault-tolerance requires that background faults be detectable and recoverable. Payment notification jobs and benchmark snapshot jobs that fail silently result in real user impact — missed invoices, stale data — with no operational signal.

Wrap every Vercel Cron route handler in try/catch with explicit logging — these routes have no automatic retry, so logging is the only failure signal.

// app/api/cron/digest/route.ts
export async function GET() {
  try {
    await sendWeeklyDigest()
    return NextResponse.json({ success: true })
  } catch (error) {
    console.error('[Cron: digest] Failed:', error)
    return NextResponse.json({ error: 'Digest failed' }, { status: 500 })
  }
}

For jobs that require reliable retry (payment processing, email delivery), migrate from Vercel Cron to Inngest or Trigger.dev — both provide automatic retry with exponential backoff and a UI for monitoring failed jobs without manual log inspection.