Retail Store en Google Cloud Platform

1

Lo que aprender谩s sobre GCP para ecommerce

2

Etapas clave y MLOps

3

Arquitectura de alto nivel

4

Tour de la aplicaci贸n de retail

5

Backend as a Service y modelo de seguridad

6

Introducci贸n al proyecto

7

Medici贸n de interacciones

8

Setup de Google Tag Manager

9

Etiquetando con Google Tag Manager

10

Etiquetas relevantes para CLV

11

Integraci贸n con servicios

Exposici贸n de servicios con Apigee

12

Servicios expuestos con APIs

13

驴Qu茅 son las APIs?

14

Apigee

15

Creaci贸n de tu primer API Proxy en Apigee

16

Administraci贸n de APIs con Apigee

17

Creando un portal de desarrolladores

18

Interactuando con el portal de desarrolladores

19

Insights to Actions

Generaci贸n de modelos AI/ML

20

Machine Learning con datos estructurados

21

BigQuery para modelos de Forecasting y LTV

22

Bigquery ML - Manos a la Obra

23

Auto ML vs. Bigquery ML

24

Consideraciones para entrenar un modelo en BigQueryML

25

Entrenamiento del modelo en BigQuery ML

26

C贸mo exportar modelos hechos en BQML

27

Exportando un modelo hecho con BQML

Consumo de servicios de AI/ML

28

C贸mputo Serverless y Contenedores

29

驴Qu茅 es Kubernetes?

30

Consumo de modelos ML mediante BigQuery API

31

Almacenamiento de predicciones

32

Ejecuci贸n de predicciones y persistencia

33

Despliegue continuo con Cloud Run

34

Ejecuci贸n de despliegue con Cloud Run

35

Escalamiento de servicios en Cloud Run

36

AuthN y AuthZ con Cloud Run

Google Marketing Platform

37

An谩lisis de las predicciones

38

Segmentamos nuestras Predicciones

39

Caso pr谩ctico para definir tu estrategia de activaci贸n

40

Generemos nuestros modelos en la plataforma

41

Segmentamos nuestras audiencias en BigQuery

42

Carga tus audiencias y conecta tu medio de activaci贸n

You don't have access to this class

Keep learning! Join and start boosting your career

Aprovecha el precio especial y haz tu profesi贸n a prueba de IA

Antes: $249

Currency
$209
Suscr铆bete

Termina en:

1 D铆as
20 Hrs
2 Min
12 Seg

Escalamiento de servicios en Cloud Run

35/42
Resources

How to prevent your application from crashing with Google Cloud Run?

In the midst of an increasingly digital world, there is nothing worse than an application crashing due to unexpectedly high traffic. Being a victim of your own success is a situation that can be avoided by properly managing the scaling options in Google Cloud Run. Let's find out how to get the most out of these tools.

What scaling options does Google Cloud Run offer?

The key to success with Cloud Run lies in understanding its two fundamentals: horizontal scaling and vertical scaling.

  1. Horizontal scaling:

    • Cloud Run offers the ability to have up to a thousand instances per service.
    • Each instance can handle up to 80 concurrent requests.
    • This means that, with the right configuration, you can handle thousands of requests simultaneously.
  2. Vertical scaling:

    • Here, each instance can be set to have between 1 and 4 CPUs.
    • Memory can be set to any value between 128 MB and 4 GB.

The big differential of Cloud Run over other serverless solutions is its ability to scale to zero, which means that you will not have instances running when not in use, which translates into cost savings.

How can we configure these options in the Google Cloud Console?

Properly configuring your scaling options is essential to overcome any potential challenges. Here's a quick guide on how to do it:

  1. Create a new service:

    • Let's go to the console, there you can create a new service named appropriately.
    • Select the image from which your service will be launched, such as one generated by Buildpack.
  2. Advanced configuration:

    • Configure the port where your application will listen, usually 8080.
    • Application startup system: define the commands as Python main.py.
    • Choose the memory between 128 MB and 4 GB. Remember that it can be customized, for example, to 345 MB.
  3. CPU and execution settings:

    • Decide between 1, 2 or 4 CPUs according to the expected load.
    • Configure the time out for long tasks, always keeping in mind that the maximum is 3600 seconds.

What are the advantages of Cloud Run over Cloud Functions?

Choosing between Cloud Run and Cloud Functions can be complicated, but Cloud Run has outstanding features:

  • Ability to handle multiple simultaneous requests: While Cloud Functions allows only one request per instance, Cloud Run can handle up to 80.
  • Flexibility in the runtime: You can customize the operating environment, which allows you to use any version of the programming languages you want to work with.
  • Cost savings: Being able to handle multiple requests in a single instance reduces costs compared to creating new instances for each request.

How to optimize scaling to get the most out of it?

The ideal configuration will depend on the type of application and specific business needs. Here are some recommendations:

  • Minimize wait times by configuring minimum instances if you are using an application with frequent use to avoid cold start.
  • Vary CPU and RAM resources to find the right balance between cost and performance, ideally try different configurations.
  • Use Environment Variables to reduce hardcoding and improve secret management with tools such as Secret Manager.

To maximize performance, it is essential to thoroughly know and understand the needs of your application. The path is to test, tweak and iterate until you find the most optimal configuration that suits your users' demands and your budget. Don't get discouraged and keep exploring new ways to improve your applications with cutting-edge technology!

Contributions 0

Questions 0

Sort by:

Want to see more contributions, questions and answers from the community?