graphql-execute
Use the graphql-execute policy to execute a GraphQL query by using sub-queries.
Gateway support
| Gateway | Policy version |
|---|---|
| DataPower® API Gateway | 2.0.0 (DataPower API Gateway Version 10.0.4.0 and later) 2.1.0 (DataPower API Gateway Version 10.5.2.0 and later) |
This topic describes how to configure the policy in your OpenAPI source; for details on how to configure the policy in the assembly user interface, see GraphQL Execute.
About
The graphql-execute policy
executes a GraphQL query by using sub-queries. The policy splits up the original GraphQL query into
different sub-queries that can then be processed in different ways. The graphql-execute policy supports the
following four different sub-queries as targets:
root; contains the core GraphQL query that you want to invoke on the backend.__introspection; contains the introspection query that can be processed locally on the DataPower API Gateway without having to invoke the backend.__rateLimit; contains the__rateLimitquery that specifies the different rate limit data that you want to receive.__cost; contains the__costquery that specifies the different cost data that you want to receive. This query must be in the format that is specified in the IBM GraphQL Cost Directives specification, see Introspection of Cost Analysis Results:__cost.
root, __rateLimit, and __introspection
targets can be executed in any order, or at the same time. The __cost target can be
executed at any time if no response data is expected. If response data is expected, then the
__cost target is executed after the root target in order to
accurately return correct response cost data. After all the targets are executed, the graphql-execute policy combines all
of the responses together. To enable the responses of the individual targets to be correctly merged
with each other into a final, single response, the graphql-execute policy must include
all of the introspection types that are associated with each level of the response data.The graphql-execute policy has
the following format:
- graphql-execute:
version: version
title: title
description: description
targets:
- execute: []
catch:
- execute: []
otherwise: []
Properties
| Property | Required | Description | Data type |
|---|---|---|---|
| version | Yes | The policy version number. | string |
| title | No | The title of the policy. | string |
| description | No | A description of the policy. | string |
| targets | No | The targets for the sub-queries. Valid values are:
|
string |
Example
The following example shows a graphql-execute policy that includes
all of the available targets.
- graphql-execute:
title: GraphQL execute
version: 2.1.0
allowCostIntrospection: false
targets:
- target: root
execute:
- validate:
input: message
version: 2.7.0
validate-against: graphql
description: >-
Validate the GraphQL query and analyze
the expected cost
title: validate & analyze
- ratelimit:
version: 2.4.0
title: ratelimit expected
description: >-
Consume expected cost of query from rate
limits
rate-limit:
- name: graphql-type-cost
operation: consume
- name: graphql-field-cost
operation: consume
burst-limit: []
source: plan-named
- invoke:
cache-ttl: 900
graphql-send-type: detect
parameter-control:
type: allowlist
values: []
http-version: HTTP/1.1
header-control:
type: allowlist
values:
- User-Agent
- X-Client-IP
- X-Global-Transaction-ID
- Content-Type
chunked-uploads: true
verb: keep
stop-on-error:
- ConnectionError
- SOAPError
- OperationError
websocket-upgrade: false
description: Invoke the backend GraphQL server
compression: false
title: graphql-invoke
target-url: >-
https://swapi-graphql.netlify.app/.netlify/functions/index
backend-type: graphql
version: 2.3.0
cache-response: protocol
timeout: 60
- parse:
title: parse
version: 2.0.0
description: >-
Parse the response from the GraphQL
backend server
parse-settings-reference:
default: apic-default-parsesettings
- validate:
input: message
version: 2.7.0
validate-against: graphql
description: >-
Validate the returned JSON and analyze
the real cost
title: validate & analyze
- ratelimit:
version: 2.4.0
title: ratelimit actual
description: >-
Replenish over-estimated cost
(determined from response) to rate
limits
rate-limit:
- name: graphql-type-cost
operation: replenish
- name: graphql-field-cost
operation: replenish
burst-limit: []
source: plan-named
catch:
- execute: []
errors:
- ConnectionError
- target: __introspection
execute:
- ratelimit:
source: plan-named
version: 2.4.0
description: >-
Apply design-time rate limit for
introspection query
title: ratelimit design-time
burst-limit: []
rate-limit:
- operation: consume
name: graphql-design-request
- validate:
input: message
version: 2.7.0
validate-against: graphql
description: >-
Validate the GraphQL query and analyze
the expected cost
title: validate & analyze
- graphql-introspect:
input: message
version: 2.0.0
description: >-
Introspect the GraphQL schema using the
query
title: graphql-introspect
- target: __rateLimit
execute:
- ratelimit:
version: 2.4.0
title: ratelimit design-time
description: >-
Apply design-time rate limit for
ratelimitinfo
rate-limit:
- name: graphql-design-request
operation: consume
burst-limit: []
source: plan-named
- validate:
version: 2.7.0
validate-against: graphql
title: validate & analyze
input: message
description: Validate the GraphQL query
- ratelimitinfo:
version: 2.0.0
title: ratelimitinfo
input: message
- target: __cost
execute:
- ratelimit:
version: 2.4.0
title: ratelimit design-time
description: >-
Apply design-time rate limit for
ratelimitinfo
rate-limit:
- name: graphql-design-request
operation: consume
burst-limit: []
source: plan-named
- validate:
version: 2.7.0
title: validate & analyze
input: message
validate-against: graphql
description: Validate the GraphQL query
- graphql-cost-analysis:
version: 2.0.0
title: graphql cost analysis
input: message
target: graphql_execute_args