Based on the CPU usage graph it looks like there's a lot more opportunity to downsize the scheduler hardware. I'd be curious to see how their TaskWorker cluster

Based on the CPU usage graph it looks like there's a lot more opportunity to downsize the scheduler hardware. I'd be curious to see how their TaskWorker cluster absorbs spikes in load when big cron jobs start.