All subgraphs come with a GraphQL interface that allows you to query the data in the subgraph. Traditionally these GraphQL interfaces are completely public and can be accessed by anyone. Goldsky supports public GraphQL endpoints for both subgraphs and their tags.

Public endpoints

For example, in the Goldsky managed community project there exists the uniswap-v3-ethereum/1.1.3 subgraph with a tag of prod. This subgraph has a public endpoint and the tag prod also has a public endpoint.

In general, public endpoints come in the form of https://api.goldsky.com/api/public/<project_id>/subgraphs/<subgraph name>/<version or tag>/gn

Goldsky adds rate limiting to all public endpoints to prevent abuse. We currently have a default rate limit of 50 requests per 10 seconds. This can be unlocked by contacting us at support@goldsky.com.

One major downside of public endpoints is that they are completely public and can be accessed by anyone. This means that anyone can query the data in the subgraph and potentially abuse the endpoint. This is why we also support private endpoints.

[BETA] Private endpoints

Private endpoints are only accessible by authenticated users. This means that you can control who can access the data in your subgraph. Private endpoints are only available to users who have been granted access to the subgraph. Accessing a private endpoint requires sending an Authorization header with the GraphQL request. The value of the Authorization header should be in the form of Bearer <token> where the token is an API token that has been generated through Goldsky project general settings. Remember that API tokens are scoped to specific projects. This means an API token for projectA cannot be used to access the private endpoints of subgraphs in projectB.

Private endpoints can be toggled on and off for each subgraph and tag. This means that you can have a mix of public and private endpoints for your subgraph. For example, you can have a public endpoint for your subgraph and a private endpoint for a specific tag.

Here’s an example of how to access a private endpoint using the GraphiQL interface:

Private subgraphs endpoints follow the same format as public subgraph endpoints except they start with /api/private instead of /api/public. For example, the private endpoint for the prod tag of the uniswap-v3-ethereum/1.1.3 subgraph would be https://api.goldsky.com/api/private/project_cl8ylkiw00krx0hvza0qw17vn/subgraphs/uniswap-v3-ethereum/prod/gn.

Revoking access

To revoke access to a private endpoint you can simply delete the API token that was used to access the endpoint. If you don’t know which key is used to access the endpoint, you’ll have to revoke all API tokens for all users that have access to the project. While this step is not ideal during this beta, this step will be addressed before this feature reaches general availability.

Enabling and disabling public and private endpoints

By default, all new subgraphs and their tags come with the public endpoint enabled and the private endpoint disabled. Both of these settings can be changed using the CLI and the webapp. To change either setting, you must have Editor permissions.

CLI

To toggle one of these settings using the CLI you can use the goldsky subgraph update command with the --public-endpoint <disabled|enabled> flag and/or the --private-endpoint <disabled|enabled> flag. Here’s a complete example disabling the public endpoint and enabling the private endpoint for the prod tag of the uniswap-v3-ethereum/1.1.3 subgraph:

goldsky subgraph update uniswap-v3-ethereum/prod --public-endpoint disabled --private-endpoint enabled

Dashboard

To toggle one of these settings using the dashboard webapp you can navigate to the subgraph detail page and use the relevant toggles to enable or disable the public or private endpoints of the subgraph or its tags.

Errors

Goldsky does not enforce CORS on our GraphQL endpoints. If you see an error that references CORS, or an error with the response code 429, you’re likely seeing an issue with rate limiting. Rate limits can be unlocked on a case-by-case basis on the Scale plan and above. Please reach out to us if you need help with rate limits or any GraphQL response errors.