In particular, when running on **Kubernetes** you will probably **not** want to use workers and instead run **a single Uvicorn process per container**, but I'll tell you about it later in that chapter.

///

## Multiple Workers { #multiple-workers }

You can start multiple workers with the `--workers` command line option:

//// tab | `fastapi`

If you use the `fastapi` command:

<div class="termy">

And if you need to have multiple workers, you can simply use the `--workers` command line option.

/// note | Technical Details

The Docker image was created when Uvicorn didn't support managing and restarting dead workers, so it was needed to use Gunicorn with Uvicorn, which added quite some complexity, just to have Gunicorn manage and restart the Uvicorn worker processes.

		t.Run(fmt.Sprintf("TestApplyNewerNoncurrentVersionsLimit-%d", i), func(t *testing.T) {
			workers := []chan expiryOp{make(chan expiryOp)}
			es.workers.Store(&workers)
			workerReady := make(chan struct{})
			var wg sync.WaitGroup
			wg.Add(1)
			var gotExpired []ObjectToDelete
			go expiryWorker(&wg, workerReady, workers[0], &gotExpired)
			<-workerReady

			item := scannerItem{
				Path:        obj,

		workers = workers[:len(workers)-1]
		worker <- expiryOp(nil)
		es.stats.workers.Add(-1)
	}
	// Atomically replace workers.
	es.workers.Store(&workers)
}

// Worker handles 4 types of expiration tasks.
// 1. Expiry of objects, includes regular and transitioned objects
// 2. Expiry of noncurrent versions due to NewerNoncurrentVersions

replication_max_workers         (number)    set the maximum number of replication workers (default: '500')
replication_max_lrg_workers     (number)    set the maximum number of replication workers MinIO uses to replicate large objects between sites. (default: '10')
transition_workers              (number)    set the number of transition workers (default: '100')

	if (checkOld > 0 && len(p.workers) != checkOld) || n == len(p.workers) || n < 1 {
		// Either already satisfied or worker count changed while we waited for the lock.
		return
	}
	for len(p.workers) < n {
		input := make(chan ReplicationWorkerOperation, 10000)
		p.workers = append(p.workers, input)

		go p.AddWorker(input, &p.activeWorkers)
	}
	for len(p.workers) > n {
		worker := p.workers[len(p.workers)-1]

But in many cases, you will want to run several worker processes at the same time.

### Multiple Processes - Workers { #multiple-processes-workers }

type NotificationGroup struct {
	workers    *workers.Workers
	errs       []NotificationPeerErr
	retryCount int
}

// WithNPeers returns a new NotificationGroup with length of errs slice upto nerrs,
// upon Wait() errors are returned collected from all tasks.
func WithNPeers(nerrs int) *NotificationGroup {
	if nerrs <= 0 {
		nerrs = 1
	}
	wk, _ := workers.New(nerrs)

// newBatchJobPool creates a pool of job manifest workers of specified size
func newBatchJobPool(ctx context.Context, o ObjectLayer, workers int) *BatchJobPool {
	jpool := &BatchJobPool{
		ctx:          ctx,
		objLayer:     o,
		jobCh:        make(chan *BatchJobRequest, 10000),
		workerKillCh: make(chan struct{}, workers),
		jobCancelers: make(map[string]context.CancelFunc),
	}

	if err != nil {
		rebalanceLogIf(ctx, fmt.Errorf("invalid workers value err: %v, defaulting to %d", err, len(pool.sets)))
		workerSize = len(pool.sets)
	}

	// Each decom worker needs one List() goroutine/worker
	// add that many extra workers.
	workerSize += len(pool.sets)

	wk, err := workers.New(workerSize)
	if err != nil {
		return err
	}