finopsspot-instancescost-optimizationarchitecture

Spot fleet diversification: capturing the discount without the drops

Spot instances are cheap but reclaimable, and concentrating a fleet in one instance type risks a mass interruption. Diversifying across types and zones smooths that risk, letting you run more of a workload on Spot safely. Here is how.

The C3X Team··4 min read

Quick answer

Spot capacity is drawn from pools defined by instance type, size, and Availability Zone. Concentrating a fleet in one pool means a single capacity shortage can reclaim everything at once. Diversifying across many instance types and zones spreads that risk, so a shortage in one pool affects only a fraction of the fleet, letting you safely run more of a workload on Spot and capture more of the discount.

Spot's discount is large but conditional: the cloud can reclaim instances when it needs the capacity. The risk that turns that from a manageable event into an outage is concentration, running your whole Spot fleet in one instance type and zone, so a single capacity shortage takes it all. Diversification is how you defuse that risk and safely run more on Spot.

Spot capacity pools

Spot capacity comes from pools, each defined by a combination of instance type, size, and Availability Zone. A shortage that triggers reclamation usually hits one pool at a time. If your fleet spans one pool, a shortage reclaims everything; if it spans twenty pools, a shortage in one affects roughly a twentieth. Diversification turns a cliff into a gentle slope.

How to diversify

DimensionDiversify across
Instance typeMultiple families and sizes that fit the workload
Availability ZoneAll zones in the region
Allocation strategyCapacity-optimized to pick pools with spare capacity
On-demand baselineA floor of on-demand for minimum capacity

Define your Spot fleet or auto-scaling group to accept many instance types and sizes that meet the workload's needs, across all zones, and use a capacity-optimized allocation strategy that draws from pools with the most spare capacity (lower reclamation risk). Keep an on-demand or reserved baseline so a broad shortage cannot drop you below a minimum, the blend from Spot performance risk.

The payoff

Diversification lets you run a larger share of a fault-tolerant workload on Spot with confidence, because no single pool shortage can take it down. That means capturing more of the up-to-90% discount than a concentrated, cautious Spot deployment would dare. Combined with graceful interruption handling and an on-demand floor, a well-diversified Spot fleet is how teams run significant production capacity on Spot safely, maximizing the saving from when Spot saves money.

FAQ

What is Spot fleet diversification?

Spreading a Spot fleet across many instance types, sizes, and Availability Zones (capacity pools) so that a shortage in any one pool reclaims only a fraction of the fleet rather than all of it. It smooths interruption risk, letting you safely run more of a workload on Spot and capture more of the discount.

Why does concentrating a Spot fleet in one instance type risk an outage?

Because Spot capacity comes from pools defined by instance type, size, and zone, and a shortage usually hits one pool at a time. A fleet concentrated in one pool can be reclaimed entirely by a single shortage, turning a manageable interruption into an outage. Diversification spreads the risk.

How do I diversify a Spot fleet?

Accept many instance types and sizes that meet the workload's needs, span all Availability Zones, use a capacity-optimized allocation strategy that draws from pools with the most spare capacity, and keep an on-demand or reserved baseline so a broad shortage cannot drop you below a minimum capacity.

What is a capacity-optimized allocation strategy?

A Spot allocation strategy that draws instances from the pools with the most available spare capacity, which have the lowest reclamation risk, rather than simply the cheapest pools. It reduces interruptions by favoring pools less likely to face a shortage, improving fleet stability while keeping the Spot discount.

Does diversification let me run more on Spot?

Yes. Because no single pool shortage can take down a well-diversified fleet, you can run a larger share of a fault-tolerant workload on Spot with confidence, capturing more of the up-to-90% discount than a concentrated, cautious deployment would risk. It is how teams run significant production capacity on Spot safely.

Does C3X estimate Spot fleet cost?

C3X prices your instances from Terraform, so you can compare on-demand and Spot pricing and model a diversified fleet with an on-demand baseline before deploy. Diversification and allocation strategy are configuration choices you make to manage interruption risk while capturing the Spot saving.

What to do next

Model your Spot and on-demand mix before you deploy. C3X reads your Terraform and prices your resources against a live catalog. Start with the quickstart.

Try C3X on your own Terraform

Free and open source. No API key required. One command to install, one command to estimate.