*/
    RDMA_WRITE,

    /**
     * Traditional send/receive with RDMA
     */
    RDMA_SEND_RECEIVE,

    /**
     * Dynamic memory registration
     */
    MEMORY_REGISTRATION,

    /**
     * Fast memory region registration
     */
    FAST_REGISTRATION

* Restarts
* Replication (the number of processes running)
* Memory
* Previous steps before starting

## Memory { #memory }

If you run **a single process per container** you will have a more or less well-defined, stable, and limited amount of memory consumed by each of those containers (more than one if they are replicated).

* Startup'ta çalıştırma
* Restart'lar
* Replication (çalışan process sayısı)
* Memory
* Başlatmadan önceki adımlar

## Memory { #memory }

**Container başına tek process** çalıştırırsanız, her container'ın tüketeceği memory miktarı aşağı yukarı tanımlı, stabil ve sınırlı olur (replication varsa birden fazla container için).

#FESS_HEAP_SIZE=2g

# Heap new generation
#FESS_HEAP_NEWSIZE=

# max direct memory
#FESS_DIRECT_SIZE=

# Additional Java OPTS
#FESS_JAVA_OPTS=

# Maximum number of open files
MAX_OPEN_FILES=65535

# Maximum amount of locked memory
#MAX_LOCKED_MEMORY=

# Fess log directory
LOG_DIR=${packaging.fess.log.dir}

# Fess data directory

You could also use it to stream **large binary files**, where you stream each chunk of data as you read it, without having to read it all in memory at once.

You could also stream **video** or **audio** this way, it could even be generated as you process and send it.

## A `StreamingResponse` with `yield` { #a-streamingresponse-with-yield }

package cmd

import (
	"errors"
	"io"

	sse "github.com/minio/minio/internal/bucket/encryption"
)

// BucketSSEConfigSys - in-memory cache of bucket encryption config
type BucketSSEConfigSys struct{}

// NewBucketSSEConfigSys - Creates an empty in-memory bucket encryption configuration cache
func NewBucketSSEConfigSys() *BucketSSEConfigSys {
	return &BucketSSEConfigSys{}
}

	processResidentMemoryBytesMD      = NewGaugeMD(processResidentMemoryBytes, "Resident memory size in bytes")
	processVirtualMemoryBytesMD       = NewGaugeMD(processVirtualMemoryBytes, "Virtual memory size in bytes")
	processVirtualMemoryMaxBytesMD    = NewGaugeMD(processVirtualMemoryMaxBytes, "Maximum virtual memory size in bytes")
)

func loadProcStatMetrics(ctx context.Context, stat procfs.ProcStat, m MetricValues) {

Y es posible que para tu caso de uso, la solución esté en una de ellas.

///

## Lee primero el Tutorial { #read-the-tutorial-first }

Aún podrías usar la mayoría de las funcionalidades en **FastAPI** con el conocimiento del [Tutorial - Guía del usuario](../tutorial/index.md) principal.

- Il utilise un fichier « spooled » :
    - Un fichier stocké en mémoire jusqu'à une taille maximale, puis, au-delà de cette limite, stocké sur le disque.
- Cela fonctionne donc bien pour des fichiers volumineux comme des images, des vidéos, de gros binaires, etc., sans consommer toute la mémoire.
- Vous pouvez obtenir des métadonnées à partir du fichier téléversé.

Esto significa que **FastAPI** fue específicamente probado con los editores usados por el 80% de los desarrolladores de Python. Y como la mayoría de los otros editores tienden a funcionar de manera similar, todos sus beneficios deberían funcionar prácticamente para todos los editores.