docs/source/using-federation/api/apollo-gateway.mdx
This API reference documents the exports from the @apollo/gateway package. This package enables you to use Apollo Server as a gateway for a federated supergraph. For more information, see Implementing a gateway with Apollo Server.
We recommend that all supergraphs use the GraphOS Router instead of this Node.js-based gateway. See Choosing a router library.
class ApolloGatewayThe core class of Apollo Server's federated gateway implementation. For an
example of using ApolloGateway, see Implementing a gateway with Apollo Server.
constructorReturns an initialized ApolloGateway instance, which you can then pass as the gateway configuration option to the ApolloServer constructor, like so:
const server = new ApolloServer({
// highlight-start
gateway: new ApolloGateway({
serviceList: [
// ...
],
}),
// highlight-end
});
Takes an options object as a parameter. Supported fields of this object are described below.
serviceList and headers to authorize introspectionconst gateway = new ApolloGateway({
serviceList: [
{ name: 'products', url: 'https://products-service.dev/graphql' },
{ name: 'reviews', url: 'https://reviews-service.dev/graphql' },
],
introspectionHeaders: {
Authorization: 'Bearer abc123',
},
});
If you provide a buildService function to the constructor of ApolloGateway, the function is called once for each of your graph's subgraphs. This function can return a RemoteGraphQLDataSource with a custom fetcher, which is then used for all communication with subgraphs:
const gateway = new ApolloGateway({
buildService({ url, name }) {
return new (class extends RemoteGraphQLDataSource {
fetcher = require('make-fetch-happen').defaults({
onRetry() {
console.log('We will retry!');
},
});
})({
url,
name,
});
},
});
<MultiCodeBlock>This only configures the fetcher that your gateway uses to fetch managed federation configuration from Apollo.
See also Configuring the subgraph fetcher.
const gateway = new ApolloGateway({
fetcher: require('make-fetch-happen').defaults({
onRetry() {
console.log('We will retry!');
},
}),
});
supergraphSdlstring | SupergraphSdlHook | SupergraphManager
You provide your supergraph schema to the gateway with this option. You can provide it as a string, via a SupergraphSdlHook, or via a SupergraphManager.
When supergraphSdl is a string: A supergraph schema (generated with the Rover CLI) that's composed from your subgraph schemas. The supergraph schema includes directives that specify routing information for each subgraph.
When supergraphSdl is a SupergraphSdlHook: This is an async function that returns an object containing a supergraphSdl string as well as a cleanup function. The hook accepts an object containing the following properties:
update: A function that updates the supergraph schemahealthCheck: A function that issues a health check against the subgraphsgetDataSource: A function that gets a data source for a particular subgraph from the gatewayWhen supergraphSdl is a SupergraphManager: An object containing an initialize property. initialize is an async function of the SupergraphSdlHook type described directly above.
If you are using managed federation, do not provide this field.
If you aren't using managed federation, either this field or serviceList is required. Do not provide both.
serviceListArray<ServiceEndpointDefinition>
This option is deprecated in favor of the drop-in replacement, IntrospectAndCompose.
An array of objects that each specify the name and url of one subgraph in your federated graph. On startup, the gateway uses this array to obtain your subgraph schemas via introspection and compose a supergraph schema.
You can specify any string value for the name field, which is used for identifying the subgraph in log output and error messages, and for reporting metrics to Apollo Studio.
If you are using managed federation, do not provide this field.
If you aren't using managed federation, either this field or supergraphSdl is required. Do not provide both.
introspectionHeadersObject | (service: ServiceEndpointDefinition) => (Promise<Object> | Object)
This option is deprecated in favor of the drop-in replacement, IntrospectAndCompose.
An object (or an optionally async function returning an object) that contains the names and values of HTTP headers that the gateway includes only when making introspection requests to your subgraphs.
If you are using managed federation, do not provide this field.
If you define a buildService function, specify these headers in that function instead of providing this option. This ensures that your buildService function doesn't inadvertently overwrite the values of any headers you provide here.
debugBoolean
If true, enables development mode helpers and logs messages of all severity levels (debug through error). If false, only warn- and error-level messages are logged.
The default value is false.
loggerAn object to use for logging in place of console. If provided, this object must implement all methods of the Logger interface.
If you provide this value, the gateway automatically logs all messages of all severity levels (debug through error), regardless of whether the debug option is set to true. It is the responsibility of the logger to determine how to handle logged messages of each level.
This logger is automatically added to the GraphQLRequestContext object that's passed to all Apollo Server plugin functions.
fetcherSpecifies which Fetch API implementation to use with managed federation when fetching configuration from Apollo.
By default, the gateway uses the default configuration of make-fetch-happen. You can specify another valid implementation (such as node-fetch) by setting this field's value to require('node-fetch').
As shown in the example above, you can also provide make-fetch-happen to this option if you want to override the library's default configuration.
serviceHealthCheckBoolean
If true, the gateway sends a small query ({ __typename }) to all subgraphs on gateway startup. It also does the same on each live schema update if you're using managed federation.
On startup, the gateway throws an error if any of these requests fail.
On schema update, the gateway does not roll over to the new schema or graph configuration if any of these requests fail. The gateway retries the requests at each polling interval until they succeed.
The default value is false.
uplinkEndpointsArray<string>
A list of Apollo Uplink URLs the gateway uses to poll for its managed configuration. For details and defaults, see Apollo Uplink.
Provide this field only if you are using managed federation and your use case specifically requires it (this is uncommon).
</td> </tr> <tr> <td>uplinkMaxRetriesnumber
The maximum number of times the gateway retries a failed poll request to Apollo Uplink, cycling through its list of uplinkEndpoints.
The default value is to try each of your uplinkEndpoints three times (i.e., 5 retries with the default list of two endpoints).
Provide this field only if you are using managed federation.
</td> </tr> <tr> <td>fallbackPollIntervalInMs (managed mode only)number
Specify this option as a fallback if Uplink fails to provide a polling interval. This will also take effect if fallbackPollIntervalInMs is greater than the Uplink defined interval.
buildServiceFunction
Define this function to customize your gateway's data transport to some or all of your subgraphs. This customization can include using a protocol besides HTTP. For details, see The buildService function.
buildService functionIf you provide this function, the gateway calls it once for each subgraph. The function must return an object that implements the GraphQLDataSource interface, such as an instance of RemoteGraphQLDataSource or a subclass of it.
The example below demonstrates adding an x-user-id HTTP header to every request the gateway sends to a subgraph:
class AuthenticatedDataSource extends RemoteGraphQLDataSource {
willSendRequest({ request, context }) {
request.http.headers.set('x-user-id', context.userId);
}
}
const gateway = new ApolloGateway({
// ...other options...
// highlight-start
buildService({ name, url }) {
return new AuthenticatedDataSource({ url });
},
// highlight-end
});
These options are experimental. They might be removed or change at any time, even within a patch release.
<table class="field-table"> <thead> <tr> <th>Name / Type</th> <th>Description</th> </tr> </thead> <tbody> <tr> <td>experimental_approximateQueryPlanStoreMiBnumber
Sets the approximate size (in MiB) of the gateway's query plan store. This cache is used to save query plans for reuse on subsequent queries that resolve to a previously observed queryHash (the SHA-256 of the incoming operation).
The default value is 30, which is usually sufficient unless the server processes a large number of unique operations.
serviceHealthCheckWhen called, the gateway sends a small query ({ __typename }) to each subgraph to verify that they're all responsive. This function throws on failure and returns a Promise to be awaited.
serviceMapObject (DataSourceMap)
If provided, the gateway sends health check requests to only the data sources included in the map.
By default, the gateway uses this.serviceMap (i.e., it sends health check requests to all known subgraphs).
class RemoteGraphQLDataSourceRepresents a connection between your federated gateway and one of your subgraphs.
You can customize this connection by extending this class and overriding its willSendRequest and/or didReceiveResponse methods:
willSendRequest to modify your gateway's requests to the subgraph before they're sent.didReceiveResponse to modify the subgraph's responses before the gateway passes them along to the requesting client.These methods are described in detail below.
constructorReturns an initialized RemoteGraphQLDataSource instance:
const productsDataSource = new RemoteGraphQLDataSource({
url: 'https://products-service.dev/graphql',
});
Takes an options object as a parameter. Supported fields of this object are described below.
urlString
Required. The subgraph's URL for sending fetch requests via HTTP.
</td> </tr> <tr> <td>apqBoolean
If true, the gateway attempts to use automated persisted queries (APQ) when sending queries to this subgraph. APQ can dramatically reduce the size of most requests sent over the network, especially for more complex queries.
The subgraph must also enable support for APQ for the gateway to use this feature (Apollo Server enables APQ by default).
</td> </tr> </tbody> </table>willSendRequestOverride this method in a subclass to modify each outgoing fetch request before it's sent to the subgraph:
<MultiCodeBlock>// Adds an `x-user-id` header to each outgoing fetch request
class AuthenticatedDataSource extends RemoteGraphQLDataSource {
willSendRequest({ request, context }) {
request.http.headers.set('x-user-id', context.userId);
}
};
This method takes a requestContext object that contains both the original unmodified request and the current context object (i.e., the contextValue object).
didReceiveResponseOverride this method in a subclass to customize the gateway's behavior after it completes a fetch to the subgraph, but before it sends a response to the requesting client:
<MultiCodeBlock>class CookieDataSource extends RemoteGraphQLDataSource {
didReceiveResponse({ response, request, context }) {
const cookie = request.http.headers.get('Cookie');
if (cookie) {
context.responseCookies.push(cookie);
}
// Return the response, even when unchanged.
return response;
}
}
This method takes a requestContext object that contains:
responserequest to the subgraphcontext (i.e., the contextValue object)This enables you to modify any combination of the operation's contextValue and the response of the fetch.
The http property of the request and response objects contains additional HTTP-specific properties, such as headers.
This method must return an object that matches the structure of a GraphQLResponse. If no modifications are necessary, return the original response.
didEncounterErrorOverride this method in a subclass to customize the gateway's behavior after it encounters an error while communicating with the subgraph or while parsing its response (e.g., if the response is not well-formed JSON).
If you don't override this method, the default implementation throws the error, like so:
class MyDataSource extends RemoteGraphQLDataSource {
didEncounterError(error, fetchRequest, fetchResponse, context) {
throw error;
}
}
Note that if you don't throw error (or a different Error) in this method, error is thrown immediately after this method returns.
This method takes the following positional parameters:
<table class="field-table"> <thead> <tr> <th>Name / Type</th> <th>Description</th> </tr> </thead> <tbody> <tr> <td>errorError
The error that occurred during communication.
</td> </tr> <tr> <td>fetchRequestRequest
The details of the fetch request sent to the subgraph.
fetchResponseResponse
The details of the fetch response sent by the subgraph.
contextTContext
The contextValue object passed between the GraphQL operation's resolvers.
class IntrospectAndCompose⚠️ We strongly recommend against using
IntrospectAndComposein production. For details, see Limitations ofIntrospectAndCompose.
IntrospectAndCompose is a development tool for fetching and composing subgraph SDL into a supergraph for your gateway. Given a list of subgraphs and their URLs, IntrospectAndCompose will issue queries for their SDL, compose them into a supergraph, and provide that supergraph to the gateway. It can also be configured to update via polling and perform subgraph health checks to ensure that supergraphs are updated safely. IntrospectAndCompose implements the SupergraphManager interface and is passed in to ApolloGateway's supergraphSdl constructor option.
IntrospectAndComposeis the drop-in replacement forserviceList.
constructorReturns an initialized IntrospectAndCompose instance, which you can then pass to the supergraphSdl configuration option of the ApolloGateway constructor, like so:
const server = new ApolloServer({
gateway: new ApolloGateway({
// highlight-start
supergraphSdl: new IntrospectAndCompose({
subgraphs: [
// ...
],
}),
// highlight-end
}),
});
Takes an options object as a parameter. Supported properties of this object are described below.
subgraphs list and headers to authorize introspectionconst gateway = new ApolloGateway({
supergraphSdl: new IntrospectAndCompose({
subgraphs: [
{ name: 'products', url: 'https://products-service.dev/graphql' },
{ name: 'reviews', url: 'https://reviews-service.dev/graphql' },
],
introspectionHeaders: {
Authorization: 'Bearer abc123',
},
}),
});
IntrospectAndCompose uses the data sources constructed by ApolloGateway. To customize the gateway's data sources, you can provide a buildService function to the ApolloGateway constructor. In the example below, IntrospectAndCompose makes authenticated requests to the subgraphs
via the AuthenticatedDataSources that we construct in the gateway's buildService function.
const gateway = new ApolloGateway({
buildService({ name, url }) {
return new AuthenticatedDataSource({ url });
},
supergraphSdl: new IntrospectAndCompose({
subgraphs: [
{ name: 'products', url: 'https://products-service.dev/graphql' },
{ name: 'reviews', url: 'https://reviews-service.dev/graphql' },
],
}),
});
subgraphsArray<ServiceEndpointDefinition>
An array of objects that each specify the name and url of one subgraph in your federated graph. On startup, IntrospectAndCompose uses this array to obtain your subgraph schemas via introspection and compose a supergraph schema.
The name field is a string that should be treated as a subgraph's unique identifier. It is used for query planning, logging, and reporting metrics to Apollo Studio.
For Studio users, subgraph names must:
introspectionHeadersObject | (service: ServiceEndpointDefinition) => (Promise<Object> | Object)
An object (or an optionally async function returning an object)that contains the names and values of HTTP headers that the gateway includes only when making introspection requests to your subgraphs.
**If you define a buildService function in your ApolloGateway config, ** specify these headers in that function instead of providing this option. This ensures that your buildService function doesn't inadvertently overwrite the values of any headers you provide here.
pollIntervalInMsnumber
Specify this option to enable supergraph updates via subgraph polling. IntrospectAndCompose polls each subgraph at the given interval.
subgraphHealthCheckboolean
This option applies only to subgraphs that are configured for polling via the
pollIntervalInMsoption. Iftrue, the gateway performs a health check on each subgraph before performing a supergraph update. Errors during health checks will result in skipping the supergraph update, but polling will continue. The health check is a simple GraphQL query (query __ApolloServiceHealthCheck__ { __typename }) to ensure that subgraphs are reachable and can successfully respond to GraphQL requests.
This option is the IntrospectAndCompose equivalent of ApolloGateway's serviceHealthCheck option. If you are using IntrospectAndCompose, enabling serviceHealthCheck on your ApolloGateway instance has no effect.
loggerAn object to use for logging in place of console. If provided, this object must implement all methods of the Logger interface.
IntrospectAndCompose doesn't share the same logger as the ApolloGateway it's configured with. In most cases, you probably want to pass the same logger to both ApolloGateway and IntrospectAndCompose.