API development lifecycle
Every organization has a unique software development lifecycle (SDLC). It is often necessary to synchronize and align API proxy deployment with the same processes that you use today for developing, testing, and deploying other applications.
API Services provides tools and RESTful APIs that enable you to integrate API proxy deployment and management into your organization's SDLC. A common usage of the RESTful API is to write scripts or code that programmatically deploy API proxies, or that migrate API proxies from one environment to another, as part of a larger automated process that also deploys or migrates other applications. API Services makes no assumptions about your SDLC (or anyone else's, for that matter). Rather, it exposes atomic functions that can be coordinated by your development team to automate and optimize your API development lifecycle.
API Services APIs are documented in the API Reference. See API reference getting started.
Every organization on Apigee Edge has at least two deployment environments that are available for API proxies: 'test' and 'prod'. The distinction between the two environments is arbitrary — each environment is simply identified by a different set of network addresses (URLs). The goal is to provide you with a domain in which you can build and verify API proxies before the API is exposed to external developers.
You can leverage these environments to synchronize API proxy development processed with your SDLC. Each environment is defined by a network address, enabling you to segregate traffic between the API proxies that you are working on, and those that are being accessed by apps at runtime. The network addresses available for each environment is defined in the set of VirtualHosts available in that environment.
Inbound, server SSL is automatically enabled for each environment. Two VirtualHosts are pre-defined in each environment:
secure. Default defines an HTTP address, while secure defines an HTTP/S address, with pre-configured server-side SSL. In an API proxy configuration, you indicate which VirtualHosts the ProxyEndpoint should listen on. When promoting to prod, you typically disable HTTP by removing the
default VirtualHost from the API proxy configuration.
For example, the following ProxyEndpoint listens on HTTP and HTTPS.
<HTTPProxyConnection> <BasePath>/v0/weather</BasePath> <Properties/> <VirtualHost>default</VirtualHost> <VirtualHost>secure</VirtualHost> </HTTPProxyConnection>
By deleting the
default VirtualHost from the ProxyEndpoint configuration, you create an API proxy that listens only on HTTPS and not on HTTP.
<HTTPProxyConnection> <BasePath>/v0/weather</BasePath> <Properties/> <VirtualHost>secure</VirtualHost> </HTTPProxyConnection>
You can see which VirtualHosts are available in an environment by selecting Environments in the management UI main menu.
Environments also provide segregation of data and resources. You can, for example, set up different caches in test and prod, which can be accessed only by API proxies executing in that environment. Additionally, API keys that are issued in the test environment are not valid in the prod environment, and vice-versa.
When you create an API proxy you'll need to decide which environment you'll be working in. You can choose to create a new API proxy on production, but that is not recommended as you may expose an API to developers before it's ready. In general you want to start by creating an API proxy in test which, after testing, you then 'promote' to prod.
Depending on your role, you may not be able to deploy to all environments. For example, the 'user' role can only deploy to the
test environment. If you're an administrator you can deploy to any environment.
For more information, see Understanding deployment.
As you work on an API proxy, API Services saves iterations of your configuration as revisions. When you deploy an API proxy, you choose a specific revision to deploy. Typically, you deploy the most recent revision, and, if necessary, revert to the previous revision number. You can choose where to deploy those revisions. For instance, you can promote a revision to prod to allow developers to start working with your API. At the same time, you may be iterating multiple revisions on test, where you're adding features or fine-tuning policies. Then, when you're ready, you can deploy the new revision to prod, overwriting the existing revision on that environment. Using this method, you can always have a live revision of your API available to developers while you're developing.
When an API proxy has been fully implemented and tested, it is ready to be promoted to 'prod'. The revision of the API proxy in test will be used to overwrite the revision of the API proxy deployed on prod.
API Services provides capabilities to ensure seamless deployment of API proxies, minimizing the impact on apps and end users during the deployment procedure.
The Apigee Edge management UI enables you to deploy API proxies to prod directly from the API proxy builder. However, in many situations the requirements for security, reliability, and consistency will mandate that development teams script deployment procedures. To do so, you can write code and scripts that invoke the RESTful API exposed by API Services.
For additional control during promotion, it is recommended that you only iterate on API proxies in test, and make as few changes as necessary to API proxies deployed in prod.
To do so, you need to ensure that certain resources associated with each environment are configured in such a way that they can remain static in an API proxy configuration.
- Target URLs: It is common for API proxies to call different backend URLs during testing and production. You can use TargetServer configurations to create environment-independent TargetEndpoint configurations. See Load balancing across backend servers.
- Caches and Key/value maps: Both persistence resources are scoped by environment. You should ensure that naming conventions are used to enable API proxies to store data without requiring configuration changes during promotion. See Manage environment caches.
- ServiceCallout targets: Service callouts may use different targets depending on the environment, if, for example, a ServiceCallout in the test environment consumes a demo service. See Service Callout policy.
To make API proxy configurations environment-independent, you can also use conditional statements. Conditional statement built with the
environment.name variable can be used to evaluate the current environment before enforcing a policy or before routing to a URL on the backend.
For more information, see Understanding deployment.