Rate limiting

Limit concurrency with batchSize

batchSize is the number of tasks that run at the same time. The next batch starts only after every task in the current batch has settled.

// At most 5 concurrent API calls
await batchProcess(userIds, async (userId) => {
  return await fetchUserDataFromAPI(userId)
}, { batchSize: 5 })

The same option is useful for keeping memory and CPU usage under control:

// Process 10 items at a time to avoid memory issues
await batchProcess(largeDataset, async (data) => {
  return await processHeavyOperation(data)
}, { batchSize: 10 })

Pace requests with delayBetweenBatches

delayBetweenBatches waits the given number of milliseconds before each batch except the first one.

// 5 requests at a time, 1 second pause between batches
await batchProcess(
  apiRequests,
  async (request) => await fetch(request.url),
  {
    batchSize: 5,
    delayBetweenBatches: 1000
  }
)

Note

Batching is a simple chunked model, not a sliding window: a slow task holds back the start of the next batch.

Splitting arrays with arrayBatch

If you only need to split an array, use arrayBatch directly:

import { arrayBatch } from 'chunk-process'

arrayBatch([1, 2, 3, 4, 5, 6, 7, 8, 9, 10], 3)
// [[1, 2, 3], [4, 5, 6], [7, 8, 9], [10]]

arrayBatch throws if batchSize is not a positive integer.