# Performance & Cost

Exoscale resources are billed by the second. The services you run should reflect what your workload genuinely needs — both now and as it evolves. Before committing to a setup, the [pricing calculator](https://calculator.exoscale.com/) lets you model any combination of services across compute, storage, databases, networking, and support plans.

## Is your setup sized for what you actually need?

The most common avoidable inefficiency on any cloud platform is capacity provisioned at launch that is never revisited. A reasonable starting point, observed under real conditions and then adjusted, is more reliable than an estimate made upfront.

Exoscale supports [scaling instances](/product/compute/instances/how-to/instance-scaling/) and [scaling DBaaS plans](/product/dbaas/how-to/scaling/) without rebuilding from scratch. For workloads with variable traffic, [Instance Pools](/product/compute/instances/how-to/instance-pools/) paired with the [Network Load Balancer](/product/networking/nlb/) let capacity follow demand automatically. For containerised workloads, [SKS node pool autoscaling](/product/compute/containers/operation/autoscaling/) does the same.

## Are you using the right storage for each purpose?

[Block Storage](/product/storage/block-storage/) is for data your instances actively read and write — OS volumes, live database files. [Object Storage (SOS)](/product/storage/object-storage/) is for everything else: backups, logs, exports, and static assets. SOS is S3-compatible and supports cross-zone [replication](/product/storage/object-storage/how-to/replication/) for resilience.

One thing worth noting: SOS objects accumulate until explicitly removed. [Lifecycle rules](/product/storage/object-storage/how-to/bucketlifecycle/) let you define retention policies that handle cleanup automatically, which is much more reliable than doing it by hand.

## Can you see how your resources are being used?

Without consistent [labels](/product/compute/instances/how-to/labels/) across your resources, attributing usage to a team, project, or environment becomes guesswork. This is straightforward to establish from the start and genuinely difficult to retrofit at scale.

[Usage Reports](/platform/usage-report/) give you per-resource consumption data. Reviewing them periodically helps catch unexpected growth early. For a more detailed look at what is running and where, the [Audit Trail](/platform/audit-trail/) records all API-level activity across your organisation.

## Going deeper

- [Pricing calculator](https://calculator.exoscale.com/) — model your setup before you build it
- [Scaling an instance](/product/compute/instances/how-to/instance-scaling/) — adjust compute capacity without rebuilding
- [DBaaS plan scaling](/product/dbaas/how-to/scaling/) — change database plan as your needs evolve
- [Instance Pools](/product/compute/instances/how-to/instance-pools/) — automatic horizontal scaling for stateless workloads
- [Block Storage](/product/storage/block-storage/) — persistent volumes for active workloads
- [Object Storage (SOS)](/product/storage/object-storage/) — durable, S3-compatible storage for backups and assets
- [SOS lifecycle rules](/product/storage/object-storage/how-to/bucketlifecycle/) — automated retention and object expiry
- [Labels](/product/compute/instances/how-to/labels/) — tagging resources for visibility and organisation
- [Usage Reports](/platform/usage-report/) — per-resource consumption data
- [Support plans](https://calculator.exoscale.com/support) — from built-in to enterprise, with defined SLAs