You're viewing Apigee Edge documentation.
Go to the
Apigee X documentation. info
What
OAuthV2 is a multi-faceted policy for performing OAuth 2.0 grant type operations. This is the primary policy used to configure OAuth 2.0 endpoints on Apigee Edge.
Tip: If you want to learn more about OAuth on Apigee Edge, see the OAuth home page. It provides links to resources, samples, videos, and more. See the advanced OAuth sample on GitHub for a good demonstration of how this policy is used in a working application.
Samples
VerifyAccessToken
VerifyAccessToken
This OAuthV2 policy configuration (with the VerifyAccessToken operation) verifies that an access token submitted to Apigee Edge is valid. When this policy operation is triggered, Edge looks for a valid access token in the request. If the access token is valid, the request is allowed to proceed. If invalid, all processing stops and an error is returned in the response.
<OAuthV2 async="false" continueOnError="false" enabled="true" name="OAuth-v20-2"> <DisplayName>OAuth v2.0 2</DisplayName> <Operation>VerifyAccessToken</Operation> <AccessTokenPrefix>Bearer</AccessTokenPrefix> <!-- Optional, default is Bearer --> </OAuthV2>
Note: Only OAuth 2.0 Bearer Tokens are supported. Message Authentication Code (MAC) tokens are not supported.
For example:
$ curl -H "Authorization: Bearer ylSkZIjbdWybfsUQe9BqP0LH5Z" http://{org_name}-test.apigee.net/weather/forecastrss?w=12797282
By default, Edge accepts access tokens in the Authorization
header with
the Bearer
prefix. You can change this default with the <AccessToken>
element.
GenerateAccessToken
Generating access tokens
For examples showing how to request access tokens for each of the supported grant types, see Requesting access tokens and authorization codes. The topic includes examples of these operations:
GenerateAuthorizationCode
Generate authorization code
For examples showing how to request authorization codes, see Requesting an authorization code.
RefreshAccessToken
Refresh an access token
For examples showing how to request access tokens using a refresh token, see Refreshing an access token.
Response flow token
Generate an access token on the response flow
Sometimes you may need to generate an access token in the response flow. For example, you may do this in response to some custom validation done in a backend service. In this example, the use case requires both an access token and a refresh token, ruling out the implicit grant type. In this case, we'll use the password grant type to generate the token. As you'll see, the trick to making this work is to pass in an Authorization request header with a JavaScript policy.
First, let's look at the sample policy:
<OAuthV2 enabled="true" continueOnError="false" async="false" name="generateAccessToken"> <Operation>GenerateAccessToken</Operation> <AppEndUser>Doe</AppEndUser> <UserName>jdoe</UserName> <PassWord>jdoe</PassWord> <GrantType>grant_type</GrantType> <ClientId>a_valid_client_id</ClientId> <SupportedGrantTypes> <GrantType>password</GrantType> </SupportedGrantTypes> </OAuthV2>
If you put this policy on the response flow, it will fail with a 401 UnAuthorized error even though the correct login parameters are specified in the policy. To solve this problem, you need to set an Authorization request header.
The Authorization header must contain a Basic access scheme with the Base64-encoded client_id:client_secret.
You can add this header with a JavaScript policy placed just before the OAuthV2 policy, like this. The "local_clientid" and "local_secret" variables must be previously set and available in the flow:
var client_id = context.getVariable("local_clientid"); var client_secret = context.getVariable("local_secret"); context.setVariable("request.header.Authorization","Basic "+CryptoJS.enc.Base64.stringify(CryptoJS.enc.Latin1 .parse(client_id + ':' + client_secret)));
See also "Encoding basic authentication credentials".
Element reference
The policy reference describes the elements and attributes of the OAuthV2 policy.
The sample policy shown below is one of many possible configurations. This sample shows an OAuthV2 policy configured for the GenerateAccessToken operation. It includes required and optional elements. Refer to the element descriptions in this section for details.
<OAuthV2 name="GenerateAccessToken"> <!-- This policy generates an OAuth 2.0 access token using the client_credentials grant type --> <Operation>GenerateAccessToken</Operation> <!-- This is in millseconds, so expire in an hour --> <ExpiresIn>3600000</ExpiresIn> <SupportedGrantTypes> <GrantType>client_credentials</GrantType> </SupportedGrantTypes> <GrantType>request.queryparam.grant_type</GrantType> <GenerateResponse/> </OAuthV2>
<OAuthV2> attributes
<OAuthV2 async="false" continueOnError="false" enabled="true" name="MyOAuthPolicy">
The following table describes attributes that are common to all policy parent elements:
Attribute | Description | Default | Presence |
---|---|---|---|
name |
The internal name of the policy. The value of the Optionally, use the |
N/A | Required |
continueOnError |
Set to Set to |
false | Optional |
enabled |
Set to Set to |
true | Optional |
async |
This attribute is deprecated. |
false | Deprecated |
<DisplayName> element
Use in addition to the name
attribute to label the policy in the
management UI proxy editor with a different, natural-language name.
<DisplayName>Policy Display Name</DisplayName>
Default |
N/A If you omit this element, the value of the policy's |
---|---|
Presence | Optional |
Type | String |
<AccessToken> element
<AccessToken>request.header.access_token</AccessToken>
By default, VerifyAccessToken expects the access token to be sent in the Authorization
header.
You can change that default using this element. For
example request.queryparam.access_token
indicates that the access token
should be present as a query parameter named access_token
.
<AccessToken>request.header.access_token</AccessToken>
is
specified:
curl https://myorg-myenv.apigee.net/oauth2/validate -H "access_token:Rft3dqrs56Blirls56a"
<AccessToken>request.queryparam.access_token</AccessToken>
is
specified:
curl "https://myorg-myenv.apigee.net/oauth2/validate?access_token:Rft3dqrs56Blirls56a"
Default: |
N/A |
Presence: |
Optional |
Type: | String |
Used with operations: |
|
<AccessTokenPrefix> element
<AccessTokenPrefix>Bearer</AccessTokenPrefix>
By default, VerifyAccessToken expects the access token to be sent in an Authorization header as a Bearer token. For example:
-H "Authorization: Bearer Rft3dqrs56Blirls56a"
Currently, Bearer is the only supported prefix.
Default: |
Bearer |
Presence: |
Optional |
Type: | String |
Valid values: |
Bearer |
Used with operations: |
|
<AppEndUser> element
<AppEndUser>request.queryparam.app_enduser</AppEndUser>
In cases where the app end user ID must be sent to the authorization server, this element lets you specify where Edge should look for the end user ID. For example, it could be sent as a query parameter or in an HTTP header.
For example request.queryparam.app_enduser
indicates that the
AppEndUser should be present as a query parameter, as, for
example, ?app_enduser=ntesla@theramin.com
. To require the AppEndUser in an HTTP
header, for example, set this value to request.header.app_enduser
.
Providing this setting enables you to include an app end user ID in the access token. This feature is useful if you want to be able to retrieve or revoke OAuth 2.0 access tokens by end user ID. For more information, see Enable retrieval and revocation of OAuth 2.0 access tokens by end user ID, app id, or both.
Default: |
N/A |
Presence: |
Optional |
Type: | String |
Valid values: |
Any flow variable accessible to the policy at runtime. |
Used with grant types: |
|
<Attributes/Attribute>
<Attributes> <Attribute name="attr_name1" ref="flow.variable" display="true|false">value1</Attribute> <Attribute name="attr_name2" ref="flow.variable" display="true|false">value2</Attribute> </Attributes>
Use this element to add custom attributes to an access token or authorization code. For example, you may wish to embed a user ID or session identifier in an access token that can be extracted and checked at runtime.
This element lets you specify a value in a flow variable or from a literal string. If you specify both a variable and a string, the value specified in the flow variable is used. If the variable cannot be resolved, then the string is the default.
For more information on using this element, see Customizing Tokens and Authorization Codes.
Displaying or hiding custom attributes in the response
Remember that if you set the GenerateResponse element of this policy to true, the full JSON representation of the token is returned in the response, including any custom attributes that you set. In some cases, you might want to hide some or all of your custom attributes in the response so that they are not visible to client apps.
By default, custom attributes do appear in the response. If you want to hide them, you can set the display parameter to false. For example:
<Attributes> <Attribute name="employee_id" ref="employee.id" display="false"/> <Attribute name="employee_name" ref="employee.name" display="false"/> </Attributes>
The value of the display
attribute is not
persisted. Let's say you generate an access token with custom attributes that you want to hide in the
generated response. Setting display=false
accomplishes that goal. However, if a new
access token is generated later using a refresh token, the original custom attributes from the
access token will show up in the refresh token response. This is because Edge does not remember that
the display
attribute was originally set to false
in the generate access token policy--the custom
attribute is simply part of the access token's metadata.
You will see the same behavior if you add custom attributes to an authorization code--when an access token is generated using that code, those custom attributes will show up in the access token response. Again, this may not be the behavior you intend.
To hide custom attributes in these cases, you have these choices:
- Explicitly reset the custom attributes in the refresh token policy and set their display to false. In this case, you may need to retrieve the original custom values from the original access token using the GetOAuthV2Info policy.
- Use a postprocessing JavaScript policy to manually extract any custom attributes that you do not want to see in the response.
See also Customizing Tokens and Authorization Codes.
Default: |
|
Presence: |
Optional |
Valid values: |
|
Used with grant types: |
|
<ClientId> element
<ClientId>request.formparam.client_id</ClientId>
In several cases, the client app must send the client ID to the authorization server. This element
specifies that Apigee should look for the client ID in the flow variable
request.formparam.client_id
. Setting ClientId
to any other variable is not supported.
See also Requesting access tokens and authorization
codes.
Default: |
request.formparam.client_id (a x-www-form-urlencoded and specified in the request body) |
Presence: |
Optional |
Type: | String |
Valid values: | The flow variable: request.formparam.client_id |
Used with grant types: |
Can also be used with the GenerateAuthorizationCode operation. |
<Code> element
<Code>request.queryparam.code</Code>
In the authorization grant type flow, the client must submit an authorization code to the authorization server (Apigee Edge). This element lets you specify where Edge should look for the authorization code. For example, it could be sent as a query parameter, HTTP header, or form parameter (the default).
The variable, request.queryparam.auth_code
indicates that the
authorization code should be present as a query parameter, as, for
example, ?auth_code=AfGlvs9
. To require the authorization code in an HTTP
header, for example, set this value to request.header.auth_code
. See also
Requesting access tokens and
authorization codes.
Default: |
request.formparam.code (a x-www-form-urlencoded and specified in the request body) |
Presence: |
optional |
Type: | String |
Valid values: | Any flow variable accessible to the policy at runtime |
Used with grant types: | authorization_code |
<ExpiresIn> element
<ExpiresIn>10000</ExpiresIn>
Enforces the expiry time of access tokens and authorization codes in milliseconds. (For
refresh tokens, use <RefreshTokenExpiresIn>
.) The expiry time value is a
system-generated value plus the <ExpiresIn>
value. If
<ExpiresIn>
is set to -1, the token or code expires as per the maximum OAuth access token expiration.
If <ExpiresIn>
is not specified, the system applies a
default value configured at the system level.
The expiry time can also be set at runtime using either a hard-coded, default value or by
referencing a flow variable. For example, you can store a token expiration value in a key value
map, retrieve it, assign it to a variable, and reference it in the policy. For example,
kvm.oauth.expires_in
.
With Apigee Edge for Public Cloud, Edge keeps the following entities in cache for a minimum of 180 seconds after the entities are accessed.
- OAuth access tokens. This means that a revoked token may still succeed for up to three minutes, until its cache limit expires.
- Key Management Service (KMS) entities (Apps, Developers, API Products).
- Custom attributes on OAuth tokens and KMS entities.
The following stanza specifies a flow variable and a default value as well. Note that the flow variable value takes precedence over the specified default value.
<ExpiresIn ref="kvm.oauth.expires_in"> 3600000 <!--default value in milliseconds--> </ExpiresIn>
Edge does not support a way to force the expiration of a token after it has been created. If you need to force token expiration (for example, based on a condition), a possible solution is described in this Apigee Community post.
By default, expired access tokens are purged from the Apigee Edge system automatically 3 days after expiration. See also Purging access tokens
Private Cloud: For an Edge for Private Cloud installation, the default
value is set by the conf_keymanagement_oauth_auth_code_expiry_time_in_millis
property.
To set this property:
- Open the
message-processor.properties
file in an editor. If the file does not exist, create it:vi /opt/apigee/customer/application/message-processor.properties
- Set the property as desired:
conf_keymanagement_oauth_auth_code_expiry_time_in_millis=3600000
- Make sure the properties file is owned by the "apigee" user:
chown apigee:apigee /opt/apigee/customer/application/message-processor.properties
- Restart the Message Processor.
/opt/apigee/apigee-service/bin/apigee-service edge-message-processor restart
Default: |
If not specified, the system applies a default value configured at the system level. |
Presence: |
Optional |
Type: | Integer |
Valid values: |
|
Used with grant types: |
Also used with GenerateAuthorizationCode operation. |
<ExternalAccessToken> element
<ExternalAccessToken>request.queryparam.external_access_token</ExternalAccessToken>
Tells Apigee Edge where to find an external access token (an access token not generated by Apigee Edge).
The variable request.queryparam.external_access_token
indicates that
the external access token should be present as a query parameter, as, for
example, ?external_access_token=12345678
. To require the external access token
in an HTTP header, for example, set this value
to request.header.external_access_token
. See also Using Third-Party OAuth
Tokens.
<ExternalAuthorization> element
<ExternalAuthorization>true</ExternalAuthorization>
If this element is false or not present, then Edge validates the client_id and client_secret normally against the Apigee Edge authorization store. Use this element when you want to work with third-party OAuth tokens. For details on using this element, see Using Third-Party OAuth Tokens.
Default: |
false |
Presence: |
Optional |
Type: | Boolean |
Valid values: | true or false |
Used with grant types: |
|
<ExternalAuthorizationCode> element
<ExternalAuthorizationCode>request.queryparam.external_auth_code</ExternalAuthorizationCode>
Tells Apigee Edge where to find an external auth code (an auth code not generated by Apigee Edge).
The variable request.queryparam.external_auth_code
indicates that
the external auth code should be present as a query parameter, as, for
example, ?external_auth_code=12345678
. To require the external auth
code
in an HTTP header, for example, set this value
to request.header.external_auth_code
. See also Using Third-Party OAuth
Tokens.
<ExternalRefreshToken> element
<ExternalRefreshToken>request.queryparam.external_refresh_token</ExternalRefreshToken>
Tells Apigee Edge where to find an external refresh token (a refresh token not generated by Apigee Edge).
The variable request.queryparam.external_refresh_token
indicates that
the external refresh token should be present as a query parameter, as, for
example, ?external_refresh_token=12345678
. To require the external refresh token
in an HTTP header, for example, set this value
to request.header.external_refresh_token
. See also Using Third-Party OAuth
Tokens.
<GenerateResponse> element
<GenerateResponse enabled='true'/>
If set to true
, the policy generates and returns a response. For example, for
GenerateAccessToken, the response might be like:
{ "issued_at" : "1467841035013", "scope" : "read", "application_name" : "e31b8d06-d538-4f6b-9fe3-8796c11dc930", "refresh_token_issued_at" : "1467841035013", "status" : "approved", "refresh_token_status" : "approved", "api_product_list" : "[Product1, nhl_product]", "expires_in" : "1799", "developer.email" : "edward@slalom.org", "token_type" : "BearerToken", "refresh_token" : "rVSmm3QaNa0xBVFbUISz1NZI15akvgLJ", "client_id" : "Adfsdvoc7KX5Gezz9le745UEql5dDmj", "access_token" : "AnoHsh2oZ6EFWF4h0KrA0gC5og3a", "organization_name" : "cerruti", "refresh_token_expires_in" : "0", "refresh_count" : "0" }
If false
, no response is sent. Instead, a set of flow variables are populated
with values related to the policy's function. For example, a flow variable called
oauthv2authcode.OAuthV2-GenerateAuthorizationCode.code
gets populated with the newly
minted auth code. Note that expires_in is expressed in seconds in the
response.
Default: |
false |
Presence: |
Optional |
Type: | string |
Valid values: | true or false |
Used with grant types: |
|
<GenerateErrorResponse> element
<GenerateErrorResponse enabled='true'/>
If set to true
, the policy generates and returns a response if the
ContinueOnError attribute is set to true. If false
(the default), no
response is sent. Instead, a set of flow variables are populated with values related to the
policy's function.
Default: |
false |
Presence: |
Optional |
Type: | string |
Valid values: | true or false |
Used with grant types: |
|
<GrantType>
<GrantType>request.queryparam.grant_type</GrantType>
Tells the policy where to find the grant type parameter that is passed in a request. Per the OAuth 2.0 specification, the grant type must be supplied with requests for access tokens and authorization codes. The variable can be a header, query parameter, or form parameter (default).
For example request.queryparam.grant_type
indicates that the Password
should be present as a query parameter, as, for example, ?grant_type=password
.
To require the grant type in an HTTP header, for example, set this value
to request.header.grant_type
. See also Requesting access tokens and authorization
codes.
Default: |
request.formparam.grant_type (a x-www-form-urlencoded and specified in the request body) |
Presence: |
Optional |
Type: | string |
Valid values: | A variable, as explained above. |
Used with grant types: |
|
<Operation> element
<Operation>GenerateAuthorizationCode</Operation>
The OAuth 2.0 operation executed by the policy.
Default: |
If |
Presence: |
Optional |
Type: | String |
Valid values: |
|
<PassWord> element
<PassWord>request.queryparam.password</PassWord>
This element is used with the password grant type only. With
the password grant type, user credentials (password and username) must be made available to the
OAuthV2 policy. The <PassWord>
and <UserName>
elements are
used to specify variables where Edge can find these values. If these elements are not specified,
the policy expects to find the values (by default) in form parameters named username
and password
. If the values are not found, the policy throws an error. You can use
the <PassWord>
and <UserName>
elements to reference any
flow variable containing the credentials.
For example, you can pass the password in a token request using a query parameter and set the
element like this:
<PassWord>request.queryparam.password</PassWord>
.
To
require the password in an HTTP header, set this value
to request.header.password
.
The OAuthV2 policy doesn't do anything else with these credential values; Edge is simply verifying that they are present. It is up to the API developer to retrieve the values request and send them to an identity provider before the token generation policy executes.
See also Requesting access tokens and authorization codes.
Default: |
request.formparam.password (a x-www-form-urlencoded and specified in the request body) |
Presence: |
Optional |
Type: | String |
Valid values: | Any flow variable available to the policy at runtime. |
Used with grant types: | password |
<RedirectUri> element
<RedirectUri>request.queryparam.redirect_uri</RedirectUri>
Specifies where Edge should look for the redirect_uri
parameter in the
request.
About redirection URIs
Redirection URIs are used with the authorization code and implicit grant types. The redirect URI tells the authorization server (Edge) where to send an authorization code (for the auth code grant type) or access token (for the implicit grant type). It's important to understand when this parameter is required, when it is optional, and how it is used:
-
(required) If a Callback URL is registered with the developer app that is associated with the request's client keys, and if the
redirect_uri
is present in the request, then the two must match exactly. If they do not match, an error is returned. For information on registering developer apps on Edge and specifying a Callback URL, see Register apps and manage API keys. - (optional) If a Callback URL is registered, and the
redirect_uri
is missing from the request, then Edge redirects to the registered Callback URL. - (required) If a Callback URL is not registered, then the
redirect_uri
is required. Note that in this case, Edge will accept ANY URL. This case can present a security issue, and therefore should only be used with trusted client apps. If client apps are not trusted, then the best practice is to always require the registration of a Callback URL.
You can send this parameter in a query parameter or in a header. The
variable request.queryparam.redirect_uri
indicates that the RedirectUri
should be present as a query parameter, as, for
example, ?redirect_uri=login.myapp.com
. To require the RedirectUri in an HTTP
header, for example, set this value to request.header.redirect_uri
. See
also Requesting access tokens and
authorization codes.
Default: |
request.formparam.redirect_uri (a x-www-form-urlencoded and specified in the request body) |
Presence: |
Optional |
Type: | String |
Valid values: | Any flow variable accessible in the policy at runtime |
Used with grant types: |
Also used with the GenerateAuthorizationCode operation. |
<RefreshToken> element
<RefreshToken>request.queryparam.refreshtoken</RefreshToken>
When requesting an access token using a refresh token, you must supply the refresh token in the request. This element lets you specify where Edge should look for the refresh token. For example, it could be sent as a query parameter, HTTP header, or form parameter (the default).
The variable request.queryparam.refreshtoken
indicates that the refresh
token should be present as a query parameter, as, for
example, ?refresh_token=login.myapp.com
. To require the RefreshToken in an HTTP
header, for example, set this value to request.header.refresh_token
. See
also Requesting access tokens and
authorization codes.
Default: |
request.formparam.refresh_token (a x-www-form-urlencoded and specified in the request body) |
Presence: |
Optional |
Type: | String |
Valid values: | Any flow variable accessible in the policy at runtime |
Used with grant types: |
|
<RefreshTokenExpiresIn> element
<RefreshTokenExpiresIn>1000</RefreshTokenExpiresIn>
Enforces the expiry time of refresh tokens in milliseconds. The expiry time value is a system
generated value plus the <RefreshTokenExpiresIn>
value. If
<RefreshTokenExpiresIn>
is set to -1, the refresh token
expires as per the maximum OAuth refresh token expiration. If <RefreshTokenExpiresIn>
is not specified,
the system applies a default value configured at the system level. Contact Apigee Edge Support for more
information about default system settings.
The expiry time can also be set at runtime using either a hard-coded, default value or by
referencing a flow variable. For example, you can store a token expiration value in a key value
map, retrieve it, assign it to a variable, and reference it in the policy. For
example, kvm.oauth.expires_in
.
The following stanza specifies a flow variable and a default value as well. Note that the flow variable value takes precedence over the specified default value.
<RefreshTokenExpiresIn ref="kvm.oauth.expires_in"> 3600000 <!--default value in milliseconds--> </RefreshTokenExpiresIn>
Private Cloud: For an Edge for Private Cloud installation, the default
value is set by the conf_keymanagement_oauth_refresh_token_expiry_time_in_millis
property.
To set this property:
- Open the
message-processor.properties
file in an editor. If the file does not exist, create it:vi /opt/apigee/customer/application/message-processor.properties
- Set the property as desired:
conf_keymanagement_oauth_refresh_token_expiry_time_in_millis=3600000
- Make sure the properties file is owned by the "apigee" user:
chown apigee:apigee /opt/apigee/customer/application/message-processor.properties
- Restart the Message Processor.
/opt/apigee/apigee-service/bin/apigee-service edge-message-processor restart
Default: |
63072000000 ms (2 years) (effective Aug 05, 2024) |
Presence: |
Optional |
Type: | Integer |
Valid values: |
|
Used with grant types: |
|
<ResponseType> element
<ResponseType>request.queryparam.response_type</ResponseType>
This element informs Edge which grant type the client app is requesting. It is used only with the authorization code and implicit grant type flows.
By default, Edge looks for the response type value in a response_type
query
parameter. If you wish to override this default behavior, use the <ResponseType> element to
configure a flow variable containing the response type value. For example, if you set this
element to request.header.response_type
, Edge looks for the response type to be
passed in the request header. See also Requesting access tokens and authorization
codes.
Default: |
request.formparam.response_type (a x-www-form-urlencoded and specified in the request body) |
Presence: |
Optional. Use this element if you wish to override the default behavior. |
Type: | String |
Valid values: | Either code (for the authorization code grant type) or token
(for the implicit grant type) |
Used with grant types: |
|
<ReuseRefreshToken> element
<ReuseRefreshToken>true</ReuseRefreshToken>
When set to true
, the existing refresh token is reused until it expires. If
false
, a new refresh token is issued by Apigee Edge when a valid refresh token is
presented.
Default: |
|
Presence: |
optional |
Type: | boolean |
Valid values: |
|
Used with grant type: |
|
<Scope> element
<Scope>request.queryparam.scope</Scope>
If this element is present in one of the GenerateAccessToken or GenerateAuthorizationCode
policies, it is used to specify which scopes to grant the token or code. These values are
typically passed to the policy in the request from a client app. You can configure the element to
take a flow variable, giving you the option to choose how the scopes are passed in a request. In
the following example, request.queryparam.scope
indicates that the scope should
be present as a query parameter, as, for example, ?scope=READ
. To require the
scope in an HTTP header, for example, set this value
to request.header.scope
.
If this element appears in a "VerifyAccessToken" policy, then it is used to specify which scopes the policy should enforce. In this type of policy, the value must be a "hard coded" scope name -- you can't use variables. For example:
<Scope>A B</Scope>
See also Working with OAuth2 scopes and Requesting access tokens and authorization codes.
Default: |
No scope |
Presence: |
Optional |
Type: | String |
Valid values: |
If used with Generate* policies, a flow variable. If used with VerifyAccessToken, a space-separated list of scope names (strings). |
Used with grant types: |
|
<State> element
<State>request.queryparam.state</State>
In cases where the client app must send the state information to the authorization server, this element lets you specify where Edge should look for the state values. For example, it could be sent as a query parameter or in an HTTP header. The state value is typically used as a security measure to prevent CSRF attacks.
For example request.queryparam.state
indicates that the state should be
present as a query parameter, as, for example, ?state=HjoiuKJH32
. To require
the state in an HTTP header, for example, set this value
to request.header.state
. See also See also Requesting access tokens and authorization
codes.
Default: |
No state |
Presence: |
Optional |
Type: | String |
Valid values: | Any flow variable accessible to the policy at runtime |
Used with grant types: |
|
<StoreToken> element
<StoreToken>true</StoreToken>
Set this element to true
when the <ExternalAuthorization>
element is true
. The <StoreToken>
element tells Apigee Edge to
store the external access token. Otherwise, it will not be persisted.
Default: |
false |
Presence: |
Optional |
Type: | Boolean |
Valid values: | true or false |
Used with grant types: |
|
<SupportedGrantTypes>/<GrantType> element
<SupportedGrantTypes> <GrantType>authorization_code</GrantType> <GrantType>client_credentials</GrantType> <GrantType>implicit</GrantType> <GrantType>password</GrantType> </SupportedGrantTypes>
Specifies the grant types supported by an OAuth token endpoint on Apigee Edge. An endpoint may
support multiple grant types (that is, a single endpoint can be configured to distribute access
tokens for multiple grant types.) For more on endpoints, see Understanding OAuth
endpoints. The grant type is passed in token requests in a grant_type
parameter.
If no supported grant types are specified, then the only allowed grant types are
authorization_code
and implicit
. See also the <GrantType> element (which is a higher-level element that is used to
specify where Apigee Edge should look for the grant_type
parameter that is passed in
a client request. Edge will make sure that the value of the grant_type
parameter
matches one of the supported grant types).
Default: |
authorization _code and implicit |
Presence: |
Required |
Type: | String |
Valid values: |
|
<Tokens>/<Token> element
Used with the ValidateToken and InvalidateToken operations. See also Approving and
revoking access tokens. The <Token> element identifies the flow variable that defines
the source of the token to be revoked. If developers are expected to submit access tokens as
query parameters named access_token
, for example,
use request.queryparam.access_token
.
<UserName> element
<UserName>request.queryparam.user_name</UserName>
This element is used with the password grant type only. With
the password grant type, user credentials (password and username) must be made available to the
OAuthV2 policy. The <PassWord>
and <UserName>
elements are
used to specify variables where Edge can find these values. If these elements are not specified,
the policy expects to find the values (by default) in form parameters named username
and password
. If the values are not found, the policy throws an error. You can use
the <PassWord>
and <UserName>
elements to reference any
flow variable containing the credentials.
For example, you can pass the username in a query parameter and set the
<UserName>
element like this:
<UserName>request.queryparam.username</UserName>
.
To require
the username in an HTTP header, set this value
to request.header.username
.
The OAuthV2 policy doesn't do anything else with these credential values; Edge is simply verifying that they are present. It is up to the API developer to retrieve the values request and send them to an identity provider before the token generation policy executes.
See also Requesting access tokens and authorization codes.
Default: |
request.formparam.username (a x-www-form-urlencoded and specified in the request body) |
Presence: |
Optional |
Type: | String |
Valid values: | Any variable setting. |
Used with grant types: | password |
Verifying access tokens
Once a token endpoint is set up for an API proxy, a corresponding OAuthV2 policy that
specifies the VerifyAccessToken
operation is attached to the Flow that exposes the
protected resource.
For example, to ensure that all requests to an API are authorized, the following policy enforces access token verification:
<OAuthV2 name="VerifyOAuthAccessToken"> <Operation>VerifyAccessToken</Operation> </OAuthV2>
The policy is attached to the API resource to be protected. To ensure that all requests to an API are verified, attach the policy to the ProxyEndpoint request PreFlow, as follows:
<PreFlow> <Request> <Step><Name>VerifyOAuthAccessToken</Name></Step> </Request> </PreFlow>
The following optional elements can be used to override the default settings for the VerifyAccessToken operation.
Name | Description |
---|---|
Scope |
A space-delimited list of scopes. Verification will succeed if at least one of the scopes listed is present in the access token. For example, the following policy will check the access token to ensure that it contains at least one of the scopes listed. If READ or WRITE is present, verification will succeed. <OAuthV2 name="ValidateOauthScopePolicy"> <Operation>VerifyAccessToken</Operation> <Scope>READ WRITE</Scope> </OAuthV2> |
AccessToken | The variable where the access token is expected to be located. For example
request.queryparam.accesstoken . By default, the access token is expected to be
presented by the app in the Authorization HTTP header, according to the OAuth 2.0 specification. Use this
setting if the access token is expected to be presented in a non-standard location, such as
a query parameter, or an HTTP header with a name other than Authorization. |
See also Verifying access tokens and Requesting access tokens and authorization codes.
Specifying request variable locations
For each grant type, the policy makes assumptions about the location or required information in request messages. These assumptions are based on the OAuth 2.0 specification. If your apps need to deviate from the OAuth 2.0 specification, then you can specify the expected locations for each parameter. For example, when handling an authorization code, you can specify the location of the authorization code, the client ID, the redirect URI, and the scope. These can be specified as HTTP headers, query parameters, or form parameters.
The example below demonstrates how you can specify the location of required authorization code parameters as HTTP headers:
... <GrantType>request.header.grant_type</GrantType> <Code>request.header.code</Code> <ClientId>request.header.client_id</ClientId> <RedirectUri>request.header.redirect_uri</RedirectUri> <Scope>request.header.scope</Scope> ...
Or, if necessary to support your client app base, you can mix and match headers and query parameters:
... <GrantType>request.header.grant_type</GrantType> <Code>request.header.code</Code> <ClientId>request.queryparam.client_id</ClientId> <RedirectUri>request.queryparam.redirect_uri</RedirectUri> <Scope>request.queryparam.scope</Scope> ...
Only one location can be configured per parameter.
Flow variables
The flow variables defined in this table are populated when the respective OAuth policies are executed, and hence are available to other policies or applications executing in the API proxy flow.
VerifyAccessToken operation
The VerifyAccessToken operation executes, a large number of flow variables are populated in the proxy's execution context. These variables give you properties related to the access token, developer app, developer, and company. You can use an AssignMessage or JavaScript policy, for example, to read any of these variables and use them as needed later in the flow. These variables can also be useful for debugging purposes.
proxy.pathsuffix
). Explicitly setting flow.resource.name variable is not required.
Where the API products are not configured with valid environments and API proxies, then
flow.resource.name
must explicitly be set to populate API product variables in the
flow. For details on product configuration, see Using the Edge management
API to Publish APIs.
Token-specific variables
Variables | Description |
---|---|
organization_name |
The name of the organization where the proxy is executing. |
developer.id |
The ID of the developer associated with the registered client app. |
developer.app.name |
The name of the developer associated with the registered client app. |
client_id |
The client ID of the registered client app. |
grant_type |
The grant type associated with the request. |
token_type |
The token type associated with the request. |
access_token |
The access token that is being verified. |
accesstoken.{custom_attribute} |
A named custom attribute in the access token. |
issued_at |
The date the access token was issued expressed in Unix epoch time in milliseconds. |
expires_in |
The expiration time for the access token. Expressed in
seconds. Although the ExpiresIn
element sets the expiration in milliseconds, in the token response and
flow variables, the value is expresed in seconds. |
status |
The status of the access token (e.g., approved or revoked). |
scope |
The scope (if any) associated with the access token. |
apiproduct.<custom_attribute_name> |
A named custom attribute of the API product associated with the registered client app. |
apiproduct.name |
The name of the API product associated with the registered client app. |
revoke_reason |
(Apigee hybrid only) Indicates why the access token is revoked. Value can be |
App-specific variables
These variables are related to the Developer App that is associated with the token.
Variables | Description |
---|---|
app.name |
|
app.id |
|
app.accessType |
|
app.callbackUrl |
|
app.status |
approved or revoked |
app.scopes |
|
app.appFamily |
|
app.apiproducts |
|
app.appParentStatus |
|
app.appType |
For example: Developer |
app.appParentId |
|
app.created_by |
|
app.created_at |
|
app.last_modified_at |
|
app.last_modified_by |
|
app.{custom_attributes} |
A named custom attribute of the registered client app. |
Developer-specific variables
If the app.appType is "Company", then company attributes are populated and if app.appType is "Developer", then developer attributes are populated.
Variables | Description |
---|---|
Developer-specific variables | |
developer.id |
|
developer.userName |
|
developer.firstName |
|
developer.lastName |
|
developer.email |
|
developer.status |
active or inactive |
developer.apps |
|
developer.created_by |
|
developer.created_at |
|
developer.last_modified_at |
|
developer.last_modified_by |
|
developer.{custom_attributes} |
A named custom attribute of the developer. |
Company-specific variables
If the app.appType is "Company", then company attributes are populated and if app.appType is "Developer", then developer attributes are populated.
Variables | Description |
---|---|
company.id |
|
company.displayName |
|
company.apps |
|
company.appOwnerStatus |
|
company.created_by |
|
company.created_at |
|
company.last_modified_at |
|
company.last_modified_by |
|
company.{custom_attributes} |
A named custom attribute of the company. |
GenerateAuthorizationCode operation
These variables are set when the GenerateAuthorizationCode operation executes successfully:
Prefix: oauthv2authcode.{policy_name}.{variable_name}
Example: oauthv2authcode.GenerateCodePolicy.code
Variable | Description |
---|---|
code |
The authorization code generated when the policy executes. |
redirect_uri |
The redirect URI associated with the registered client app. |
scope |
The optional OAuth scope passed in the client request. |
client_id |
The client ID passed in the client request. |
GenerateAccessToken and RefreshAccessToken operations
These variables are set when the GenerateAccessToken and RefreshAccessToken operations execute successfully. Note that refresh token variables do not apply for the client credentials grant type flow.
Prefix: oauthv2accesstoken.{policy_name}.{variable_name}
Example: oauthv2accesstoken.GenerateTokenPolicy.access_token
Variable name | Description |
---|---|
access_token |
The access token that was generated. |
client_id |
The client ID of the developer app associated with this token. |
expires_in |
The expiry value for the token. See the <ExpiresIn> element for details. Note that in the response, expires_in is expressed in seconds. |
scope |
List of available scopes configured for the token. See Working with OAuth2 scopes. |
status |
Either approved or revoked . |
token_type |
Is set to BearerToken . |
developer.email |
The email address of the registered developer who owns the developer app associated with the token. |
organization_name |
The org where the proxy executes. |
api_product_list |
A list of the products associated with the token's corresponding developer app. |
refresh_count |
|
refresh_token |
The refresh token that was generated. Note that refresh tokens are not generated for the client credentials grant type. |
refresh_token_expires_in |
The lifespan of the refresh token, in seconds. |
refresh_token_issued_at |
This time value is the string representation of the corresponding 32-bit timestamp quantity. For example, 'Wed, 21 Aug 2013 19:16:47 UTC' corresponds to the timestamp value of 1377112607413. |
refresh_token_status |
Either approved or revoked . |
GenerateAccessTokenImplicitGrant
These variables are set when the GenerateAccessTokenImplicit operation executes successfully for the implicit grant type flow.
Prefix: oauthv2accesstoken.{policy_name}.{variable_name}
Example: oauthv2accesstoken.RefreshTokenPolicy.access_token
Variable | Description |
---|---|
oauthv2accesstoken.access_token |
The access token generated when the policy executes. |
oauthv2accesstoken.{policy_name}.expires_in |
The expiry value for the token, in seconds. See the <ExpiresIn> element for details. |
Error reference
This section describes the fault codes and error messages that are returned and fault variables that are set by Edge when this policy triggers an error. This information is important to know if you are developing fault rules to handle faults. To learn more, see What you need to know about policy errors and Handling faults.
Runtime errors
These errors can occur when the policy executes.
Fault code | HTTP status | Cause | Thrown by operations |
---|---|---|---|
steps.oauth.v2.access_token_expired |
401 | The access token is expired. |
VerifyAccessToken |
steps.oauth.v2.access_token_not_approved |
401 | The access token was revoked. | VerifyAccessToken |
steps.oauth.v2.apiresource_doesnot_exist |
401 | The requested resource does not exist any of the API products associated with the access token. | VerifyAccessToken |
steps.oauth.v2.FailedToResolveAccessToken |
500 | The policy expected to find an access token in a variable specified in the
<AccessToken> element, but the variable could not be resolved. |
GenerateAccessToken |
steps.oauth.v2.FailedToResolveAuthorizationCode |
500 | The policy expected to find an authorization code in a variable specified in the
<Code> element, but the variable could not be resolved. |
GenerateAuthorizationCode |
steps.oauth.v2.FailedToResolveClientId |
500 | The policy expected to find the Client ID in a variable specified in the
<ClientId> element, but the variable could not be resolved. |
GenerateAccessToken GenerateAuthorizationCode GenerateAccessTokenImplicitGrant RefreshAccessToken |
steps.oauth.v2.FailedToResolveRefreshToken |
500 | The policy expected to find a refresh token in a variable specified in the
<RefreshToken> element, but the variable could not be resolved. |
RefreshAccessToken |
steps.oauth.v2.FailedToResolveToken |
500 | The policy expected to find a token in a variable specified in the
<Tokens> element, but the variable could not be resolved. |
ValidateToken |
steps.oauth.v2.InsufficientScope |
403 | The access token presented in the request has a scope that does not match the scope specified in the verify access token policy. To learn about scope, see Working with OAuth2 scopes. | VerifyAccessToken |
steps.oauth.v2.invalid_access_token |
401 | The access token sent from the client is invalid. | VerifyAccessToken |
steps.oauth.v2.invalid_client |
401 |
This error name is returned when the Note: It is recommended that you change existing fault rule
conditions to catch both the |
GenerateAccessToken RefreshAccessToken |
steps.oauth.v2.invalid_request |
400 | This error name is used for multiple different kinds of errors, typically for missing
or incorrect parameters sent in the request. If <GenerateResponse> is
set to false , use fault variables (described below) to retrieve details about
the error, such as the fault name and cause. |
GenerateAccessToken GenerateAuthorizationCode GenerateAccessTokenImplicitGrant RefreshAccessToken |
steps.oauth.v2.InvalidAccessToken |
401 | The authorization header does not have the word "Bearer", which is required. For
example: Authorization: Bearer your_access_token |
VerifyAccessToken |
steps.oauth.v2.InvalidAPICallAsNo\ |
401 |
The API proxy is not in the Product associated with the access token. Tips: Be sure that the product associated with the access token is configured correctly. For example, if you use wildcards in resource paths, be sure the wildcards are being used correctly. See Create API products for details. See also this Apigee Community post for more guidance on causes for this error. |
VerifyAccessToken |
steps.oauth.v2.InvalidClientIdentifier |
500 |
This error name is returned when the |
GenerateAccessToken |
steps.oauth.v2.InvalidParameter |
500 | The policy must specify either an access token or an authorization code, but not both. | GenerateAuthorizationCode GenerateAccessTokenImplicitGrant |
steps.oauth.v2.InvalidTokenType |
500 | The <Tokens>/<Token> element requires you to specify the token
type (for example, refreshtoken ). If the client passes the wrong type, this
error is thrown. |
ValidateToken InvalidateToken |
steps.oauth.v2.MissingParameter |
500 | The response type is token , but no grant types are specified. |
GenerateAuthorizationCode GenerateAccessTokenImplicitGrant |
steps.oauth.v2.UnSupportedGrantType |
500 |
The client specified a grant type that is unsupported by the policy (not listed in the <SupportedGrantTypes> element). Note: There is currently a bug where unsupported grant type errors are not thrown correctly. If an unsupported grant type error occurs, the proxy does not enter the Error Flow, as expected. |
GenerateAccessToken GenerateAuthorizationCode GenerateAccessTokenImplicitGrant RefreshAccessToken |
Deployment errors
These errors can occur when you deploy a proxy containing this policy.
Error name | Cause |
---|---|
InvalidValueForExpiresIn |
For the |
InvalidValueForRefreshTokenExpiresIn |
For the <RefreshTokenExpiresIn> element, valid values are positive
integers and -1 . |
InvalidGrantType |
An invalid grant type is specified in the <SupportedGrantTypes>
element. See the policy reference for a list of valid types. |
ExpiresInNotApplicableForOperation |
Be sure that the operations specified in the <Operations> element support expiration. For example, the VerifyToken operation does not. |
RefreshTokenExpiresInNotApplicableForOperation |
Be sure that the operations specified in the <Operations> element support refresh token expiration. For example, the VerifyToken operation does not. |
GrantTypesNotApplicableForOperation |
Be sure that the grant types specified in <SupportedGrantTypes> are supported for the specified operation. |
OperationRequired |
You must specify an operation in this policy using the Note: If the |
InvalidOperation |
You must specify a valid operation in this policy using the
Note: If the |
TokenValueRequired |
You must specify a token <Token> value in the
<Tokens> element. |
Fault variables
These variables are set when this policy triggers an error at runtime.
<GenerateResponse>
is set to
false
. If <GenerateResponse>
is
true
, the policy returns a response to the client immediately if an error occurs --
the error flow is skipped and these variables are not populated. For more information, see
What you need to
know about policy errors.Variables | Where | Example |
---|---|---|
fault.name="fault_name" |
fault_name is the name of the fault, as listed in the Runtime errors table above. The fault name is the last part of the fault code. | fault.name = "invalid_request" |
oauthV2.policy_name.failed |
policy_name is the user-specified name of the policy that threw the fault. | oauthV2.GenerateAccesstoken.failed = true |
oauthV2.policy_name.fault.name |
policy_name is the user-specified name of the policy that threw the fault. | oauthV2.GenerateAccesstoken.fault.name = invalid_request
Note: For the VerifyAccessToken operation, the fault name includes this suffix: |
oauthV2.policy_name.fault.cause |
policy_name is the user-specified name of the policy that threw the fault. | oauthV2.GenerateAccesstoken.cause = Required param : grant_type |
Example error response
These responses are sent back to the client if the <GenerateResponse>
element is true.
errorcode
part of the error response. Do not rely on the text in the
faultstring
, because it could change.If <GenerateResponse>
is true, the policy returns errors
in this format for operations that generate tokens and codes. For a complete list, see see
OAuth HTTP error
response reference.
{"ErrorCode" : "invalid_client", "Error" :"ClientId is Invalid"}
If <GenerateResponse>
is true, the policy returns errors
in this format for verify and validate operations. For a complete list, see see OAuth HTTP error
response reference.
{ { "fault":{ "faultstring":"Invalid Access Token", "detail":{ "errorcode":"keymanagement.service.invalid_access_token" } } }
Example fault rule
<FaultRule name=OAuthV2 Faults"> <Step> <Name>AM-InvalidClientResponse</Name> <Condition>(fault.name = "invalid_client") OR (fault.name = "InvalidClientIdentifier")</Condition> </Step> <Step> <Name>AM-InvalidTokenResponse</Name> <Condition>(fault.name = "invalid_access_token")</Condition> </Step> <Condition>(oauthV2.failed = true) </Condition> </FaultRule>
Policy schema
Each policy type is defined by an XML schema (.xsd
). For reference, policy
schemas are available on GitHub.
Working with the default OAuth configuration
Each organization (even a free trial org) on Apigee Edge is provisioned with an OAuth token endpoint. The endpoint is preconfigured with policies in the API proxy called oauth. You can begin using the token endpoint as soon as you create an account on Apigee Edge. For details, see Understanding OAuth endpoints.
Purging access tokens
By default, OAuth2 tokens are purged from the Apigee Edge system 3 days (259200 seconds) after both the access token and refresh token (if it exists) have expired. In some cases, you may want to change this default. For example, you may want to shorten the purge time to save disk space if a large number of tokens are being generated.
If you are on Edge for Private Cloud, you can change this default by setting organization properties as explained in this section. (The 3-day purge of expired tokens applies to Edge for Private Cloud version 4.19.01 and later. For earlier versions, the default purge interval is 180 days.)
Updating purge settings for Edge Private Cloud 4.16.01 and later versions
Note: Only tokens generated after these settings are applied are affected; the settings do not apply to tokens that were generated earlier.
- Open this file for editing:
/opt/apigee/customer/application/message-processor.properties
- Add the following property to set the number of seconds to wait before purging a token
after it expires:
conf_keymanagement_oauth.access.token.purge.after.seconds=<number of seconds>
- Restart the message processor. For example:
/opt/apigee/apigee-service/bin/apigee-service edge-message-processor restart
<ExpiresIn>
and
<RefreshTokenExpiresIn>
attributes.
Updating purge settings for Edge Private Cloud 4.15.07
Note: Only tokens generated after these settings are applied are affected; the settings do not apply to tokens that were generated earlier.
-
Set positive values for the
<ExpiresIn>
and<RefreshTokenExpiresIn>
attributes in the OAuthV2 policy. Values are in milliseconds. For example:<ExpiresIn>1000</ExpiresIn> <RefreshTokenExpiresIn>10000</RefreshTokenExpiresIn>
-
Redeploy the proxy.
-
Use this API to update the token purge properties for your organization:
POST https://<host-name>/v1/organizations/<org-name>
Payload:
<Organization name="AutomationOrganization"> <Description>Desc</Description> <Properties> <Property name="keymanagement.oauth20.access.token.purge">true</Property> <Property name="keymanagement.oauth20.access.token.purge.after.seconds">120</Property> </Properties> </Organization>
-
Restart the message processor. For example:
/opt/apigee/apigee-service/bin/apigee-service edge-message-processor restart
This API sets the token purge property to true for the organization called AutomationOrganization. In this case, the access token will be purged from the database 120 seconds after both the token and refresh token expire.
Non-RFC-compliant behavior
The OAuthV2 policy returns a token response that contains certain non-RFC-compliant properties. The following table shows the non-compliant properties returned by the OAuthV2 policy and the corresponding compliant properties.
OAuthV2 returns: | The RFC-compliant property is: |
---|---|
"token_type":"BearerToken" |
"token_type":"Bearer"
|
"expires_in":"3600" |
"expires_in":3600
(The compliant value is a number, not a string.) |
Also, the error response for an expired refresh token when grant_type = refresh_token
is:
{"ErrorCode" : "invalid_request", "Error" :"Refresh Token expired"}
However, the RFC-compliant response is:
{"error" : "invalid_grant", "error_description" :"refresh token expired"}