Cost Optimizer spotted $18 / month in idle compute, your 4 production Droplets still need backups, and llama-3 inference is running 60% slower than it could on a dedicated SFO3 instance.
$18 / mo
Right-size 2 idle Droplets
web-dev-1 + web-dev-2 averaged 4% CPU over the last 30 days.
Why this?Cost Optimizer flagged sustained < 5% CPU on production-tagged droplets — high confidence the workload fits a smaller plan.