CloudRun
Overview
Versus GKE
Oliver Fierro - Cloud Architect
https://www.linkedin.com/in/oliverfierro/
oliverfierro77@gmail.com
Definition
Cloudrun is a serverless product in GCP that offer the capabilities to execute
applications over a container without the need to manage the infrastructure or the
cluster for your own.
Run stateless containers on a fully managed environment.
Running Mode
● CloudRun
● CloudRun for Anthos
Advantages
Automatic scale
Pay for use
Automanaged platform
Language Support
Go, Python, Java, C#, PHP, Ruby, Node.js, Shell, and others
Cost
Pay for application execution time
CloudRun versus GKE
CLOUDRUN GKE
Service YML Internal Recetas kubernetes.
Source Code Config Cloudbuild.yaml o Cloudbuild.json deployment.yaml
Secrets No. Environment Var at config level.
Sensible data into Secret Manager.
Environment var. Sensitive values encrypted (secrets).
Exposing Services Dominio https Differents kind of service exposition (LoadBalancer, NodePort,
ClusterIP)
Logging Stackdriver logging Stackdriver logging
Monitoring Metrics Dashboard Stackdriver monitoring GKE
Image pulling Container Registry or Artifact Registry Service account like a secrets to pull image from private registry.
CloudRun versus GKE
CLOUDRUN GKE
Unsupported GCP
Products
Memorystore - Filestore (NAS) - Cloud Load Balancing
Google Cloud Armor - Cloud CDN - Identity-Aware Proxy
VPC Service Controls - Cloud Asset Inventory
Infrastructure
Management
Fully managed
Serverless
Manage configuration and cluster (scaling, memory, cpu, fine tuning,
pod’s, service exposure)
Application
inter-communication
Synchronous: Https call
Asynchronous: Cloud Tasks or PubSub
Between pod exposed internally (tcp/udp load balancer) and/or
externally (http load balancer)
Deploy strategies Rollbacks, partial deploy, traffic management Recreate, RollingUpdate, Canary, A/B, Blue/Green
Cloud Native Microservicios desplegados en contenedores como servicio Microservicios dentro de contenedores desplegados en el cluster
Pipeline [cloudbuild.yml]
Upload image from gitlab source to cloud repository: gcloud builds
command
Deploying via console
[service.yml, deployment.yml, kustomization.yml]
Soporte para YML (deployment, service)
Ecosistema de microservicios intercomunicados
Deploying using k8s (terraform)
CloudRun metrics
CloudRun logging
CloudRun details
CloudRun yaml
CloudRun permissions
Others Configurations
Build and Deploy Image
Build:
$ gcloud builds submit --config cloudbuild.yaml
Deploy:
$ gcloud run deploy --image gcr.io/[PROJECT-ID]/[IMAGE_NAME]--platform managed
OR
Console: Select Image from Container Registry Image
CloudRun
Overview
Versus GKE
Oliver Fierro - Cloud Architect
https://www.linkedin.com/in/oliverfierro/
oliverfierro77@gmail.com

GCP CloudRun Overview

  • 1.
    CloudRun Overview Versus GKE Oliver Fierro- Cloud Architect https://www.linkedin.com/in/oliverfierro/ oliverfierro77@gmail.com
  • 2.
    Definition Cloudrun is aserverless product in GCP that offer the capabilities to execute applications over a container without the need to manage the infrastructure or the cluster for your own. Run stateless containers on a fully managed environment.
  • 3.
    Running Mode ● CloudRun ●CloudRun for Anthos
  • 4.
    Advantages Automatic scale Pay foruse Automanaged platform
  • 5.
    Language Support Go, Python,Java, C#, PHP, Ruby, Node.js, Shell, and others
  • 6.
  • 7.
    CloudRun versus GKE CLOUDRUNGKE Service YML Internal Recetas kubernetes. Source Code Config Cloudbuild.yaml o Cloudbuild.json deployment.yaml Secrets No. Environment Var at config level. Sensible data into Secret Manager. Environment var. Sensitive values encrypted (secrets). Exposing Services Dominio https Differents kind of service exposition (LoadBalancer, NodePort, ClusterIP) Logging Stackdriver logging Stackdriver logging Monitoring Metrics Dashboard Stackdriver monitoring GKE Image pulling Container Registry or Artifact Registry Service account like a secrets to pull image from private registry.
  • 8.
    CloudRun versus GKE CLOUDRUNGKE Unsupported GCP Products Memorystore - Filestore (NAS) - Cloud Load Balancing Google Cloud Armor - Cloud CDN - Identity-Aware Proxy VPC Service Controls - Cloud Asset Inventory Infrastructure Management Fully managed Serverless Manage configuration and cluster (scaling, memory, cpu, fine tuning, pod’s, service exposure) Application inter-communication Synchronous: Https call Asynchronous: Cloud Tasks or PubSub Between pod exposed internally (tcp/udp load balancer) and/or externally (http load balancer) Deploy strategies Rollbacks, partial deploy, traffic management Recreate, RollingUpdate, Canary, A/B, Blue/Green Cloud Native Microservicios desplegados en contenedores como servicio Microservicios dentro de contenedores desplegados en el cluster Pipeline [cloudbuild.yml] Upload image from gitlab source to cloud repository: gcloud builds command Deploying via console [service.yml, deployment.yml, kustomization.yml] Soporte para YML (deployment, service) Ecosistema de microservicios intercomunicados Deploying using k8s (terraform)
  • 9.
  • 10.
  • 11.
  • 12.
  • 13.
  • 14.
  • 15.
    Build and DeployImage Build: $ gcloud builds submit --config cloudbuild.yaml Deploy: $ gcloud run deploy --image gcr.io/[PROJECT-ID]/[IMAGE_NAME]--platform managed OR Console: Select Image from Container Registry Image
  • 16.
    CloudRun Overview Versus GKE Oliver Fierro- Cloud Architect https://www.linkedin.com/in/oliverfierro/ oliverfierro77@gmail.com