Create a high-quality deployment process

Supplying and using APIs creates dependencies. An API Service-Level Agreement (SLA) describes the marginal conditions that apply while the API is in service. You should note that the service level of each API in use is subject to customer expectations, even if no service-level agreement had been specified.

Incorrect API management can constitute a significant barrier to its use. The deployment and use of an open API is almost always restricted in some way.

In practice, the API service level is managed using the API Gateway. It often sets rules and terms relating to the service level. For example, the total number of queries can be limited, or some user queries are prioritised. It is important to inform API users about such rules and terms to ensure that their expectations regarding API operation are realistic.

If there is no communication about the service level of an API, using it is a risk that the user cannot assess. From the user's point of view, the API could in theory stop working any time.

In a good deployment process, any restrictions are communicated about and justified openly

• to those planning to use the API
• as part of the deployment process
• as part of the service-level agreement while in use.

A practical example of a poorly implemented API deployment process is one in which the access request is sent on a paper form to an agency, which can go on processing it for months. When estimating the processing time is not possible, an important project may be delayed.

Regulation may naturally set marginal conditions on the extent to which a process can be automated.

You should aim for clear communication about and description of the process.

Consider

• what kind of customer experience the user will have
• what the deployment process looks like to the user.