docs: add a central glossary

Signed-off-by: Aramis <sennyeyaramis@gmail.com>
This commit is contained in:
Aramis
2024-01-25 23:42:14 -05:00
parent 9a8d556661
commit 2e440d4723
11 changed files with 341 additions and 97 deletions
+4 -10
View File
@@ -4,22 +4,16 @@ title: Concepts
description: A list of important permission framework concepts
---
### Permission
Any action that a user performs within Backstage may be represented as a permission. More complex actions, like executing a software template, may require authorization for multiple permissions throughout the flow. Permissions are identified by a unique name and optionally include a set of attributes that describe the corresponding action. Plugins are responsible for defining and exposing the permissions they enforce.
### Policy
User permissions are authorized by a central, user-defined permission policy. At a high level, a policy is a function that receives a Backstage user and permission, and returns a decision to allow or deny. Policies are expressed as code, which decouples the framework from any particular authorization model, like role-based access control (RBAC) or attribute-based access control (ABAC).
### Policy decision versus enforcement
Two important responsibilities of any authorization system are to decide if a user can do something, and to enforce that decision. In the Backstage permission framework, policies are responsible for decisions and plugins (typically backends) are responsible for enforcing them.
### Resources and rules
In many cases, a permission represents a user's interaction with another object. This object likely has information that policy authors can use to define more granular access. The permission framework introduces two abstractions to account for this: resources and rules. Resources represent the objects that users interact with. Rules are predicate-based controls that tap into a resource's data. For example, the catalog plugin defines a resource for catalog entities and a rule to check if an entity has a given annotation.
In many cases, a permission represents a user's interaction with another object. This object likely has information that policy authors can use to define more granular access. The permission framework introduces two abstractions to account for this: [resources](../references/glossary.md#permission-resource) and [rules](../references/glossary.md#permission-rule). For example, the catalog plugin defines a resource for catalog entities and a rule to check if an entity has a given annotation.
### Conditional decisions
Rules need additional data before they can be used in a decision. For example, the catalog plugin's "has annotation" rule needs to know what annotation to look for on a given entity. Once a rule is bound to relevant information it forms a condition. Conditions are then used to return a conditional decision from a policy. Conditional decisions tell the permission framework to delegate evaluation to the plugin that owns the corresponding resource. Permission requests that result in a conditional decision are allowed if all of the provided conditions evaluate to be true. This conditional behavior avoids coupling between policies and resource schemas, and allows plugins to evaluate complex rules in an efficient way. For example, a plugin may convert a conditional decision to a database query instead of loading and filtering objects in memory.
See [Conditional decisions](../references/glossary.md#conditional-decisions).
A good example would be the catalog plugin's "has annotation" rule needs to know what annotation to look for on a given entity. The permission framework would respond to a request by the catalog plugin in this case with a condition decision. The catalog plugin would then need to correctly filter for entities matching the "has annotations" condition. This conditional behavior avoids coupling between policies and resource schemas, and allows plugins to evaluate complex rules in an efficient way. For example, a plugin may convert a conditional decision to a database query instead of loading and filtering objects in memory.
+1 -1
View File
@@ -4,7 +4,7 @@ title: Defining custom permission rules
description: How to define custom permission rules for existing resources
---
For some use cases, you may want to define custom [rules](./concepts.md#resources-and-rules) in addition to the ones provided by a plugin. In the [previous section](./writing-a-policy.md) we used the `isEntityOwner` rule to control access for catalog entities. Let's extend this policy with a custom rule that checks what [system](https://backstage.io/docs/features/software-catalog/system-model#system) an entity is part of.
For some use cases, you may want to define custom [rules](../references/glossary.md#permission-rules) in addition to the ones provided by a plugin. In the [previous section](./writing-a-policy.md) we used the `isEntityOwner` rule to control access for catalog entities. Let's extend this policy with a custom rule that checks what [system](https://backstage.io/docs/features/software-catalog/system-model#system) an entity is part of.
## Define a custom rule
@@ -4,9 +4,9 @@ title: 2. Adding a basic permission check
description: Explains how to add a basic permission check to a Backstage plugin
---
If the outcome of a permission check doesn't need to change for different [resources](../concepts.md#resources-and-rules), you can use a _basic permission check_. For this kind of check, we simply need to define a [permission](../concepts.md#resources-and-rules), and call `authorize` with it.
If the outcome of a permission check doesn't need to change for different [resources](../../references/glossary#permission-resource), you can use a _basic permission check_. For this kind of check, we simply need to define a [permission](../../references/glossary.md#permission), and call `authorize` with it.
For this tutorial, we'll use a basic permission check to authorize the `create` endpoint in our todo-backend. This will allow Backstage integrators to control whether each of their users is authorized to create todos by adjusting their [permission policy](../concepts.md#policy).
For this tutorial, we'll use a basic permission check to authorize the `create` endpoint in our todo-backend. This will allow Backstage integrators to control whether each of their users is authorized to create todos by adjusting their [permission policy](../../references/glossary.md#policy).
We'll start by creating a new permission, and then we'll use the permission api to call `authorize` with it during todo creation.
@@ -4,7 +4,7 @@ title: 3. Adding a resource permission check
description: Explains how to add a resource permission check to a Backstage plugin
---
When performing updates (or other operations) on specific [resources](../concepts.md#resources-and-rules), the permissions framework allows for the decision to be based on characteristics of the resource itself. This means that it's possible to write policies that (for example) allow the operation for users that own a resource, and deny the operation otherwise.
When performing updates (or other operations) on specific [resources](../../references/glossary.md#permission-resource), the permissions framework allows for the decision to be based on characteristics of the resource itself. This means that it's possible to write policies that (for example) allow the operation for users that own a resource, and deny the operation otherwise.
## Creating the update permission