Lambda

We’re overhauling Dgraph’s docs to make them clearer and more approachable. If you notice any issues during this transition or have suggestions, please let us know.

Introduction

For Shared and Free backends, a Dgraph lambda script is uniquely identified by a deployment id associated with the backend. You can identify your backend and the associated deployment id by using the List Backends API. In case of a Dedicated, and multi-tenant backend, an additional key, the tenant id is also required to uniquely identify a lambda script. This tenant id key is ignored when used in the context of Shared and Free backends.

As a first step, you will need to identify your backend.

List Deployment and Get Lambda Script

Use the List Backends API to identify your backend, as well as get the lambda script deployed. In order to list the backends, you will need to pass a Bearer token as an Authorization header. This token is generated by logging in via the Login query first.

Cloud Endpoint

https://cerebro.cloud.dgraph.io/graphql

API Command

query {
  deployments {
    uid
    name
    zone
    subdomain
    url
    tenantID
    lambdaScript
  }
}

Arguments

None

Example

#!/usr/bin/env bash

CEREBRO_JWT="<cerebro-jwt-from-login-query>"

curl "https://cerebro.cloud.dgraph.io/graphql" \
  -H 'Content-Type: application/json' \
  -H "Authorization: Bearer ${CEREBRO_JWT}" \
  --data-binary '{"query":"query{\n  deployments {\n    uid\n    name\n    zone\n    subdomain\n    url\n    tenantID\n    lambdaScript\n  }\n}","variables":{}}' \
  --compressed

The uid field in the response of the query is to be used as the deployment id. In the above example, the deployment id for the backend with name “OrderJourney” is “0x6238”. The field lambdaScript contains the lambda script in the form of a Base64 encoded string.

Decode the Base64 encoded lambdaScript

In order to decode the Base64 encoded string into the actual lambda code, please use the command as shown below.

$ echo "Ly8gWW91IGNhbiB0eXBlL3Bhc3RlIHlvdXIgc2NyaXB0IGhlcmUKY29uc3QgTmFtZVJlc29sdmVyID0gKHtwYXJlbnQ6IHtuYW1lfX0pID0+IGBNeSBuYW1lIGlzICR7bmFtZX0uYAoKc2VsZi5hZGRHcmFwaFFMUmVzb2x2ZXJzKHsKICAgICJQZXJzb24ubmFtZSI6IE5hbWVSZXNvbHZlcgp9KQ==" | base64 -d

Output

// You can type/paste your script here
const NameResolver = ({ parent: { name } }) => `My name is ${name}.`

self.addGraphQLResolvers({
  "Person.name": NameResolver,
})

Lambda Logs

You can fetch the logs for your lambda by using the getLambdaLogs query.

Cloud Endpoint

https://cerebro.cloud.dgraph.io/graphql

API Command

query GetLambdaLogs($lambdaLogsInput: LambdaLogsInput!) {
  getLambdaLogs(input: $lambdaLogsInput)
}

Arguments

• lambdaLogsInput: a LambdaLogsInput object
• lambdaLogsInput.deploymentID: the deployment UID returned from [List Backends]((./backend#list-backends)
• lambdaLogsInput.tenantID: In case of a multi-tenant, and dedicated backend, you will need to pass the tenant Id as well
• lambdaLogsInput.start: start time
• lambdaLogsInput.end: end time

Example

#!/usr/bin/env bash

CEREBRO_JWT="<cerebro-jwt-from-login-query>"

curl "https://cerebro.cloud.dgraph.io/graphql" \
  -H 'Content-Type: application/json' \
  -H "Authorization: Bearer ${CEREBRO_JWT}" \
  --data-binary '{"query":"query GetLambdaLogs($input: LambdaLogsInput!) {\n getLambdaLogs(input: $input)\n}","variables":{"input":{"deploymentID":"0xf0ffe9"}}}' \
  --compressed

Update or Delete Lambda

You can update or delete your lambda by using the updateLambda mutation. You will need to pass a deployment id to uniquely identify your lambda. In case your backend is multi-tenant, you will need to pass the tenant Id as well.

In order to update your lambda, you would need to convert your lamdba script into a Base64 encoded format and send it as the lambdaScript argument. In order to delete your lambda, you can simply send in an empty string in the lambdaScript argument.

Cloud Endpoint

https://cerebro.cloud.dgraph.io/graphql

API Command

mutation updateLambda($input: UpdateLambdaInput!) {
  updateLambda(input: $input)
}

Arguments

• updateLambdaInput: an UpdateLambdaInput object
• updateLambdaInput.deploymentID: the deployment UID returned from [List Backends]((./backend#list-backends)
• updateLambdaInput.tenantID: the tenant ID in case your backend is a dedicated, and multi-tenant backend. In case you have any other type of backend, you can leave it as the default 0 value
• updateLambdaInput.lambdaScript: the base64-encoded Javascript string containing your Lambda Resolver

Example

1. Create your Lambda Resolver script

//your lambda resolver

2. Base64 encode your script

$ echo "//your lambda resolver" | base64

Ly95b3VyIGxhbWJkYSByZXNvbHZlcgo=

3. Send in a cURL request

#!/usr/bin/env bash

CEREBRO_JWT="<cerebro-jwt-from-login-query>"

curl "https://cerebro.cloud.dgraph.io/graphql" \
  -H 'Content-Type: application/json' \
  -H "Authorization: Bearer ${CEREBRO_JWT}" \
  --data-binary '{"query":"mutation updateLambda($input: UpdateLambdaInput!){\n  updateLambda(input: $input)\n}","variables":{"input":{"deploymentID":"0x6238","tenantID":0,"lambdaScript":"Ly95b3VyIGxhbWJkYSByZXNvbHZlcgo="}}}' \
  --compressed