Legacy scopes
What is a scope?
These scopes are legacy. They remain supported for existing workloads, but for new implementations we recommend using agent-backed scopes instead.
See Agent-backed scopes for more information.
Scopes are runtime environments where your application assets are deployed. A scope represents a logical unit where the asset (e.g., Docker image or Lambda function) runs and scales with the same technological specification.
In a server-based application, you can also think of a scope as a cluster or pool of VMs that run replicas of your application. In a serverless application, a scope is the host of a single function.
What kinds of scopes exist?
We support Kubernetes, Server instances (AWS EC2), and Serverless (AWS Lambda).
How do I partition scopes? How many can I have?
Scopes can be used for different purposes such as:
- Distribute an application across multiple regions,
- Create production, testing, and development environments
- Separate reads from writes,
- Rightsize your infrastructure by allocating different hardware to different workloads,
- ... and more
You can have as many scopes as you want, so apply and combine the criteria above to create what best fits your use case.
On AWS, each balancer supports up to 100 deployments. There's a hard-limit of 100 deployments per balancer
(specifically per ELBListener) imposed by AWS, so you should take care of distributing scopes across different
balancers to avoid reaching this limit. See the Increasing Scope Limit in AWS section
Management for more details about
this subject.
Can anybody create or modify scopes?
Scopes are subject to approvals and policies. Refer to those topics to learn how to manage scope security in your organization.