This. If I had too many jobs running on underpowered single point of failure hardware, I wouldn't immediately think cron is my problem and rewrite it. Rethink your architecture, write meaningful log messages, collect stats, figure out why so many of your services fail that often.