Using tanstack query for async task management

This stack overflow post talks about the when to use the mutation hook vs the useQuery hook. The gist of it was:

• If you're reading data: useQuery
• If you're creating/updating/deleting data: useMutation

Is there a good approach/solution for workflows that involve creating a task, polling its status and fetching its results (in my example these are 3 separate endpoints)?

The useQuery hook seems to be a nightmare for this because it often involves:

• a mutationFn to start the process
• separate useQuery hooks for polling and fetching results
• enabling/disabling each hook based on some state variables
• extra code and hooks outside of the function to update the application with the data returned
• The builtin refetch and retry are a plus here though

The mutation hook isn't as bad:

• I shove all the logic into the one mutationFn
• I have to write the polling logic and the retry logic manually

Does others that use this library have an approach to this?

Here's the gist of what I do today

mutationFn: async () => {
  const data = await startTask();
  let status = await getStatus(data);

  // I'll need to manually add some retry logic here
  while (status === "processing") {
    await sleep(1000);
    status = await getStatus(data);
  }
  const results = await getResults(data);
  return results;
}