A GraphQL Rate Limiter to add basic but granular rate limiting to your Queries or Mutations.
- πββοΈ Add rate limits to queries or mutations
- π€ Works with any Node.js GraphQL setup (@directive, graphql-shield rule and a base rate limiter function for every other use case)
- π Add filters to rate limits based on the query or mutation args
- β Custom error messaging
- β° Configure using a simple
max
perwindow
arguments - πΌ Custom stores, use Redis, Postgres, Mongo... it defaults to in-memory
- πͺ Written in TypeScript
yarn add graphql-rate-limit
import { createRateLimitDirective } from 'graphql-rate-limit';
// Step 1: get rate limit directive instance
const rateLimitDirective = createRateLimitDirective({ identifyContext: (ctx) => ctx.id });
const schema = makeExecutableSchema({
schemaDirectives: {
rateLimit: rateLimitDirective
},
resolvers: {
Query: {
getItems: () => [{ id: '1' }]
}
},
typeDefs: gql`
directive @rateLimit(
max: Int,
window: String,
message: String,
identityArgs: [String],
arrayLengthField: String
) on FIELD_DEFINITION
type Query {
# Step 2: Apply the rate limit instance to the field with config
getItems: [Item] @rateLimit(window: "1s", max: 5, message: "You are doing that too often.")
}
`
});
import { createRateLimitRule } from 'graphql-rate-limit';
// Step 1: get rate limit shield instance rule
const rateLimitRule = createRateLimitRule({ identifyContext: (ctx) => ctx.id });
const permissions = shield({
Query: {
// Step 2: Apply the rate limit rule instance to the field with config
getItems: rateLimitRule({ window: "1s", max: 5 })
}
});
const schema = applyMiddleware(
makeExecutableSchema({
typeDefs: gql`
type Query {
getItems: [Item]
}
`,
resolvers: {
Query: {
getItems: () => [{ id: '1' }]
}
}
}),
permissions
)
import { getGraphQLRateLimiter } from 'graphql-rate-limit';
// Step 1: get rate limit directive instance
const rateLimiter = getGraphQLRateLimiter({ identifyContext: (ctx) => ctx.id });
const schema = makeExecutableSchema({
typeDefs: `
type Query {
getItems: [Item]
}
`,
resolvers: {
Query: {
getItems: async (parent, args, context, info) => {
// Step 2: Apply the rate limit logic instance to the field with config
const errorMessage = await rateLimiter(
{ parent, args, context, info },
{ max: 5, window: '10s' }
);
if (errorMessage) throw new Error(errorMessage);
return [{ id: '1' }]
}
}
}
})
You'll notice that each usage example has two steps, step 1 we get an instace of a rate limiter and step 2 we apply the rate limit to one or more fields. When creating the initial instance we pass 'Instance Config' (e.g. identifyContext
or a store
instance), this instance will likely be the only instance you'd create for your entire GraphQL backend and can be applied to multiple fields.
Once you have your rate limiting instance you'll apply it to all the fields that require rate limiting, at this point you'll pass field level rate limiting config (e.g. window
and max
).
And so... we have the same 'Instance Config' and 'Field Config' options which ever way you use this library.
A required key and used to identify the user/client. The most likely cases are either using the context's request.ip, or the user ID on the context. A function that accepts the context and returns a string that is used to identify the user.
identifyContext: (ctx) => ctx.user.id
An optional key as it defaults to an InMemoryStore. See the implementation of InMemoryStore if you'd like to implement your own with your own database.
store: new MyCustomStore()
Generate a custom error message. Note that the message
passed in to the field config will be used if its set.
formatError: ({ fieldName }) => `Woah there, you are doing way too much ${fieldName}`
Generate a custom error. By default, a RateLimitError
instance is created when a request is blocked. To return an instance of a different error class, you can return your own error using this field.
createError: (message: string) => new ApolloError(message, '429');
This enables a per-request synchronous cache to properly rate limit batch queries. Defaults to false
to preserve backwards compatibility.
enableBatchRequestCache: false | true
Specify a time interval window that the max
number of requests can access the field. We use Zeit's ms
to parse the window
arg, docs here.
Define the max number of calls to the given field per window
.
If you wanted to limit the requests to a field per id, per user, use identityArgs
to define how the request should be identified. For example you'd provide just ["id"]
if you wanted to rate limit the access to a field by id
. We use Lodash's get
to access nested identity args, docs here.
A custom message per field. Note you can also use formatError
to customise the default error message if you don't want to define a single message per rate limited field.
Limit calls to the field, using the length of the array as the number of calls to the field.
It is recommended to use a persistent store rather than the default InMemoryStore. GraphQLRateLimit currently supports Redis as an alternative. You'll need to install Redis in your project first.
import { createRateLimitDirective, RedisStore } from 'graphql-rate-limit';
const GraphQLRateLimit = createRateLimitDirective({
identifyContext: ctx => ctx.user.id,
/**
* Import the class from graphql-rate-limit and pass in an instance of redis client to the constructor
*/
store: new RedisStore(redis.createClient())
});