# Deployment targets

Locations that software may be deployed to and run can include:
- Containers running in clusters
- Cloud-managed services, platforms, or serverless functions
- Virtual machines or infrastructure as a service
- Self-managed or managed servers
- Point-of-sale devices in retail stores
- Medical devices in hospitals

All these places are different kinds of deployment targets. 

A deployment target is a location that will host your software. We've used the term _deployment target_ as this could refer to many different destinations, such as:
- Kubernetes clusters
- Cloud apps or services
- Cloud storage
- Windows or Linux servers
- On-premises machines
- Serverless functions
- SSH connections

Deployment targets define where your software is deployed, and they also serve as the basis for Day-2 tasks and additional operational tasks that can be managed by Octopus.

Deployments to Kubernetes clusters are performed by a lightweight agent that runs on the cluster. If you’re deploying to a Windows or Linux server or virtual machine, your deployment target will run a lightweight agent called a Tentacle. For cloud services, such as Amazon ECS or Azure Web Apps, the deployment is made through a worker instead.

You select the target type when you add it to Octopus. Based on the type, you'll be prompted to set up the appropriate connection using a simple form.

## Learn more
- [How Octopus counts deployment targets](/docs/infrastructure/deployment-targets/how-octopus-counts-deployment-targets)
- [Adding deployment targets](/docs/getting-started/first-deployment/add-deployment-targets)