CostMonStart free

Cloud cost

On-demand vs. spot instances

On-demand pricing charges the full published rate with no commitment. Spot pricing sells unused capacity at a steep discount, but the provider can reclaim it on short notice.

Cloud cost

Last updated

Definition

On-demand is the default way to pay for compute. It costs the most per hour. You make no commitment, get no discount, and pay only for what you use. Spot capacity is the provider's unused inventory. The discount can reach up to 90 percent off on-demand. The trade-off is that the provider can reclaim a spot instance with as little as two minutes' warning.

Spot fits stateless work that tolerates interruption. Examples are batch jobs, CI runners, and fault-tolerant worker fleets. It is a poor fit for anything that cannot checkpoint or restart cleanly.

Where it shows up

On the bill, on-demand and spot usage of the same instance type appear as separate line items. Each has its own published rate. In the console, a spot request carries its own status field. AWS sends an interruption notice a short time before it reclaims that capacity.

What makes it expensive

Can this workload survive losing its instance mid-run? Answer that before you move anything stateful onto spot capacity. On spot, interruption is routine, not rare. A fleet that starts absorbing more interruptions has not become less suited to spot. It was never suited, and the discount was hiding the cost of every restart.

One flat price. Read-only access.

CostMon charges a flat rate, not a cut of what you spend or save. Access is read-only, so nothing you connect can be changed from our side.

Esc