The Discovery agent finds deployed API Proxies in Apigee then sends them to API Central

The Apigee Discovery Agent is built and distributed as a docker image. To execute the agent using the image, the following command may be executed. Update the version in the examples below.

docker run --env-file env_vars -v `pwd`/keys:/keys ghcr.io/axway/apigee_discovery_agent:v0.1.31

If using a local specs path, mount into the /specs volume and set the APIGEE_SPECCONFIG_LOCALPATH variable to /specs.

docker run --env-file env_vars -v `pwd`/specs:/specs -v `pwd`/keys:/keys ghcr.io/axway/apigee_discovery_agent:v0.1.31

This is the default operating mode that discoveries API Proxies and attempts to match them to Specs

• Find all specs (unless disabled, see options below)
  • Parse all specs to determine endpoints with in
  • Save info to cache
• Find all Deployed API Proxies
  • Find the Spec
    • If local specs path set, see options below, check for the spec there using the Proxy Name as the file name and searching using the extensions
    • Proxy Revision has spec set, use it
    • Proxy Revision has association.json resource file, get path
      • Using path check to see if it is in the specs that were found by agent, use it
    • Using deployed URL path check for specs for match, use it
  • Check proxy for Key or Oauth policy for authentication
  • Create API Service
    • If the spec was found, use it in revision
    • If the spec was not found, create as unstructured, given option to do so is set (see below)
    • Attach appropriate Credential Request Definition based on policy in proxy

• Managed Application
  • Creates a new App on Apigee under the configured developer
• Access Request
  • Creates a new Product, or uses existing, based off the APIGEE-Proxy and Central Plan combination
  • Associates the new Product to any existing Credentials on the Application
• Credential
  • Creates a new Credential on the App and associates all Access Requests products to it

This mode can be setting the APIGEE_DISCOVERYMODE environment variable to product

• Find all specs
  • Parsing within the specs job is not necessary in this mode
  • Save info to cache
  • If using a local specs path, set the "spec_local" attribute with a value of the spec file name in that path
• Find all Products defined
  • Using the APIGEE_FILTER determine if the product should be discovered
  • Determine spec
    • If a spec_local attribute is set on the product look for the spec in the local specs path
    • Using the product's name or display name, match it to a spec (case insensitive)
  • If a spec is found, create an API Service (create as unstructured when no spec is found, if optino set)
    • Use product definition, add attributes to Service
    • Donwload and attach spec file

• Managed Application
  • Creates a new App on Apigee under the configured developer
• Access Request
  • Creates a new Product, or uses existing, using the product associated with the API Service as a template
  • Associates the new Product to any existing Credentials on the Application
• Credential
  • Creates a new Credential on the App and associates all Access Requests products to it

In both modes the provisioning process will set quota values on the created Product when handling Access Requests. In order for Apigee to enforce quota based on the values set in teh Product a Quota Enforcement Policy needs to be set on the deployed Proxy.

Here is a sample Quota policy that may be added to the desired Proxies.

<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
<Quota async="false" continueOnError="false" enabled="true" name="impose-quota">
    <DisplayName>Impose Quota</DisplayName>
    <Synchronous>true</Synchronous>
    <Distributed>true</Distributed>
    <Identifier ref="developer.app.name"/>
    <Allow countRef="verifyapikey.Verify-API-Key-1.apiproduct.developer.quota.limit"/>
    <Interval ref="verifyapikey.Verify-API-Key-1.apiproduct.developer.quota.interval "/>
    <TimeUnit ref="verifyapikey.Verify-API-Key-1.apiproduct.developer.quota.timeunit"/>
</Quota>

• Display Name - the name of the policy
• Distributed and Synchronous - set to true to ensure the counter is updated and enforced in a distributed and synchronous manner. Setting to false may allow extra calls.
• Identifier - how the proxy will count usage across multiple apps, so each get their own quota
• Allow - in this case using the API Key policy gets the quota limit from the product definition
• Interval - in this case using the API Key policy gets the quota interval from the product definition
• TimeUnit - in this case using the API Key policy gets the quota time unit from the product definition Í

The following make targets are available

make docker-build

docker run --env-file env_vars -v `pwd`/keys:/keys apigee-discovery-agent:latest

• Build the agent using the following command

go build -tags static_all \
    -ldflags="-X 'github.com/Axway/agent-sdk/pkg/cmd.BuildTime=$${time}' \
        -X 'github.com/Axway/agent-sdk/pkg/cmd.BuildVersion=$${version}' \
        -X 'github.com/Axway/agent-sdk/pkg/cmd.BuildCommitSha=$${commit_id}' \
        -X 'github.com/Axway/agent-sdk/pkg/cmd.BuildAgentName=APIGEEDiscoveryAgent'" \
    -a -o ./bin/apigee_discovery_agent.exe ./main.go

• After a successful build, you should see the executable under the bin folder. And you can execute it using the following command

./apigee_discovery_agent.exe --envFile env_vars