如何为Apollo单个Mutation延长超时并禁止重试？

Hey there, let's tackle your timeout and duplicate upload problems head-on. Here's how to adjust your Apollo Client setup to handle large 10MB JSON uploads properly:

1. Set a Custom Timeout for a Single Mutation #

You can specify a longer timeout directly in the useMutation hook using the context option. This lets you override any global timeout settings just for this specific operation.

Here's how to implement it:

const [uploadLargeJson] = useMutation(YOUR_UPLOAD_MUTATION, {
  context: {
    // Set timeout to 60 seconds (adjust based on your needs)
    fetchOptions: { timeout: 60000 }
  },
  // Add your existing onCompleted/onError handlers here
});

This timeout value is in milliseconds, so 60000 equals 1 minute. This applies only to this mutation, leaving other operations using your default timeout settings.

2. Disable Retries for the Specific Mutation #

Your current RetryLink is set to retry up to 5 times for any error, which is likely causing the duplicate uploads when the timeout hits. To stop this, we'll add a custom flag to the mutation's context and update the RetryLink to respect that flag.

First, modify your RetryLink configuration:

const retryLink = new RetryLink({
  delay: { initial: 300, max: Infinity, jitter: true },
  attempts: { 
    max: 5, 
    retryIf: (error, operation) => {
      // Check if the mutation has a "skipRetry" flag in its context
      const shouldSkipRetry = operation.getContext().skipRetry;
      // Only retry if there's an error AND the flag isn't set
      return !!error && !shouldSkipRetry;
    }
  },
});

Then, add the skipRetry flag to your useMutation call:

const [uploadLargeJson] = useMutation(YOUR_UPLOAD_MUTATION, {
  context: {
    fetchOptions: { timeout: 60000 },
    skipRetry: true // This tells RetryLink to skip retries for this operation
  },
});

Now this mutation won't retry even if it hits a timeout or other error, preventing duplicate data from being sent to your database.

A Quick Note on Server-Side Timeouts #

Don't forget to check your server's timeout settings too! If your backend or reverse proxy (like Nginx) has a shorter timeout than what you set on the client, it might still terminate the request early. Make sure those values are aligned with your client-side timeout to avoid unexpected issues.

内容的提问来源于stack exchange，提问作者Kevin Danikowski