Cost Optimization
Understand Edge's pricing model and practical ways to optimise spend without sacrificing performance.
Edge's pricing model
Edge uses per-resource pricing: compute is billed hourly for exactly the vCPU, RAM, and disk you configure, while DNS is free and CDN and Storage have free tiers. Crucially, egress is zero — you don't pay for bandwidth between your origin and the CDN, or from the CDN to users. That removes a major variable cost found on traditional clouds.
Compute runs from a prepaid balance: you top up, usage is deducted hourly, and spending can never exceed what you've added. A surprise bill is impossible by design — the cost lever is simply how much you provision.
Right-sizing VMs
Don't over-provision. A small static site or API may run fine on 1 vCPU / 1 GB RAM. Check your VM's actual CPU and memory usage in the control panel under Compute → Metrics.
If utilisation is consistently low (e.g. under 20%), resize to a smaller configuration. Edge lets you resize VMs without data loss — and because billing is per resource, dropping a vCPU or halving RAM cuts your hourly cost immediately. Scaling up is just as easy when you need it.
Monitoring your spend
The Billing page in the control panel shows your prepaid balance, hourly burn rate, and a breakdown of what each resource costs. The Usage page tracks vCPU, RAM, storage, CDN requests, and bandwidth over time.
Auto top-up: enable it so your balance refills automatically before it runs out — the top-up amount scales with your usage, and running VMs are stopped if the balance ever reaches zero, so it's worth keeping headroom.
When to optimise: if your hourly rate feels high, right-size VMs, enable CDN caching to reduce origin load, and clean up unused storage, snapshots, and idle machines.
CDN caching and storage lifecycle
CDN caching reduces origin load — fewer requests hit your VM, so you use fewer compute resources. Configure cache rules for static assets (long TTL) and bypass for dynamic content. See our Caching Strategies guide.
Storage lifecycle: Clean up unused files. Delete old backups, test data, and temporary uploads. Use the storage dashboard to identify large buckets or paths. If you use versioning, consider lifecycle policies to expire old versions — check the Storage docs for supported options.
Stopped isn't free — but it's cheap
Stopping a VM stops the vCPU and RAM charges, but its disk (and any snapshots) keep billing while they exist — you're paying for the reserved storage. For dev machines you use occasionally, stopping between sessions cuts most of the cost. For machines you're done with, delete them and their snapshots to stop the meter entirely.
The Usage page
Make the Usage page a regular stop. It breaks down consumption by resource type and shows trends over time. Use it to spot growth before you hit limits and to validate that optimisation changes (e.g. caching, resizing) are having the expected effect.
Set up alerts if available — we notify you when you approach plan limits so you can act before overages occur.