alphagov / data-standards-authority

Collaboration space for working on data standards and guidance for the DSA
https://alphagov.github.io/data-standards-authority/
Other
24 stars 11 forks source link

Feedback on 'APIM - 01 - Introduction' #118

Open jamietanna opened 2 years ago

jamietanna commented 2 years ago

On https://alphagov.github.io/data-standards-authority/guidance/apimanagement/APIM-Introduction.html, some minor content suggestions:

-If you have any APIs that need access management, monitoring or documentation, you should at the very least use an API gateway. Even a single API will benefit from the standardised tools that an API gateway provides, by not having to develop and deploy those elements of functionality. If you later build more APIs, having an API gateway in place already will mean you don’t have to build those tools again, reducing duplication of effort for future teams.
+If you have any APIs that need access management, monitoring or documentation, you should at the very least use an API gateway. Even a single API will benefit from the standardised tools that an API gateway provides, as you will not have to develop and deploy those elements of functionality. If you later build more APIs, having an API gateway in place already will mean you don’t have to build those tools again, reducing duplication of effort for future teams.
-There will be cases where your current ecosystem doesn’t fit with this model, perhaps because you already have APIs in production which have their own security or monitoring, or because the APIs have been developed using disparate methods and technologies, in which case you should think about how you can put in place standardised patterns. This is about processes and governance that teams will use to build and deploy APIs, rather than the specific technology they use. As an example, you should consider how prescriptive you want your publishing pipeline to be. You might decide to have strict rules only for certain things like security, while allowing other areas to be more flexible.
+There will be cases where your current ecosystem doesn’t fit with this model, perhaps because you already have APIs in production which have their own security or monitoring, or because the APIs have been developed using disparate methods and technologies, in which case you should think about how you can put in place standardised patterns. This is about processes and governance that teams will use to build and deploy APIs, rather than the specific technology they use.
+
+As an example, you should consider how prescriptive you want your publishing pipeline to be. You might decide to have strict rules only for certain things like security, while allowing other areas to be more flexible.
-For example, it is common for a central team to run the API Gateway, and therefore have control over service levels and capacity. You should then consider how that responsibility might interact with the design and delivery responsibilities of an API owner, such as versioning.
+For example, it is common for a central team to run the API Gateway, and therefore have control over service levels and capacity. You should then consider how that responsibility might interact with the design and delivery responsibilities of an API owner, such as versioning or security.