Google Cloud Run is a fully managed serverless platform that allows you to deploy and run containerized applications without worrying about infrastructure. Whether you’re running a small API or a large-scale production app, Cloud Run scales seamlessly based on incoming requests.
🔑 Key Features of Cloud Run Services
🌍 Unique HTTPS Endpoint
Every service deployed on Cloud Run gets a unique HTTPS endpoint. This means your service is accessible securely out of the box, with no need to configure certificates manually.
🔒 Private and Public Services
- You can expose your services publicly to the internet.
- Or keep them private within your organization for internal workloads.
- This flexibility makes it perfect for both customer-facing apps and backend microservices.
💰 Pay-per-Use Pricing
Forget about paying for idle resources. With Cloud Run, you are billed only for the exact compute time and resources used while handling requests.
👉 This makes it cost-efficient compared to traditional VM-based deployments.
🔄 Built-in Traffic Management
Cloud Run comes with native traffic splitting and version management.
- Deploy a new version and gradually shift traffic (e.g., 10% → 50% → 100%).
- Roll back instantly if issues occur.
This is incredibly useful for A/B testing or canary deployments.
⚡ Fast, Request-Based Auto Scaling
Cloud Run automatically scales your containers based on incoming requests.
- Zero instances when idle → Scale down to 0.
- Thousands of instances during traffic peaks → Scale horizontally in seconds.
- This ensures you only pay for what you need, with the ability to handle sudden traffic spikes effortlessly.
✅ Why Choose Cloud Run?
- Developer friendly – Deploy with just one command.
- Secure by default – HTTPS, IAM integration, private networking.
- Scalable and cost-effective – From zero to thousands of requests instantly.
To Create a Google Cloud Run Services
Step-01: Introduction
- Create a Cloud Run Service
- Update Applications
- Revision URLs
- Traffic Splitting
- Autoscaling
- Implement all the above features using gcloud run
Step-02: Create Service and Access it
- Go to Cloud Run -> Create Service
- Deploy one revision from an existing container image: stacksimplify/google-cloud-run:v1
- Service Name: myservice1
- Authentication: Allow unauthenticated invocations
- Container port: 80
- REST ALL LEAVE TO DEFAULTS
- Click on CREATE
# Docker Image used
stacksimplify/google-cloud-run:v1
# Access Application on Browser
https://myservice1-506997606680.us-central1.run.app/
Google Cloud Run Services – HTTPS Endpoint
Unique HTTPS endpoint for every service
- Each service with unique sub domain *.run.app
- Supports custom domains
- Manages TLS
- Support for WebSockets, HTTP/2 and gRPC (both end-to-end).
Google Cloud Run Services – Pricing Model
- Pay-per-use pricing for services
- Request-based
If an instance is not processing requests, the CPU is not allocated and not charged. Additionally, we pay a per-request fee.
- Instance-based
You’re charged for the entire lifetime of an instance and the CPU is always allocated. There’s no per-request fee.
Step-03: Update Application – v2
- Go to Cloud Run -> myservice1 -> EDIT & DEPLOY NEW REVISION
- Deploy one revision from an existing container image:
stacksimplify/google-cloud-run:v2
- Serve this revision immediately: CHECKED
- Click on DEPLOY
# Docker Image used
stacksimplify/google-cloud-run:v2
# Access Application on Browser
https://myservice1-506997606680.us-central1.run.app/
Google Cloud Run Services – Traffic Management
Built-in traffic management – Route Traffic to
- Latest revision
- Roll back to previous revision
- Split Traffic to multiple revisions at same time (gradual rollout)
Step-04: Cloud Run Revisions and Traffic Splitting
- Split Traffic between version 1 and version 2
- version-1: 50%
- version-2: 50%
# Access Application on Browser
https://myservice1-506997606680.us-central1.run.app/
Step-05: Add Revision URLs
- Add Revision URLs
- version-1: myappv1
- version-2: myappv2
# myappv1 Revision URL
https://myappv1---myservice1-czbx2i66ca-uc.a.run.app/
# myappv2 Revision URL
https://myappv2---myservice1-czbx2i66ca-uc.a.run.app/
Step-06: Deploy V3 Application with Serve this revision immediately UNCHECKED
- Go to Cloud Run -> myservice1 -> EDIT & DEPLOY NEW REVISION
- Deploy one revision from an existing container image:
stacksimplify/google-cloud-run:v3
- Serve this revision immediately: UNCHECKED
- Click on DEPLOY
# Docker Image used
stacksimplify/google-cloud-run:v3
# Access Application on Browser
https://myservice1-czbx2i66ca-uc.a.run.app
Observation: V2 version will be still serving
Add Revision URL for V3
- version-3: myappv3
# Access Application on Browser using Revision URL
https://myappv3---myservice1-czbx2i66ca-uc.a.run.app/
Observation:
1. V2 version will be still serving
2. V3 is started serving after revision url added
Traffic Split to V3: 10%
# Access Application on Browser using Revision URL
https://myservice1-czbx2i66ca-uc.a.run.app/
Observation:
1. V1 version will be serving 40%
2. v2 version will be serving 50%
2. V3 version will be serving 10% - Gradual Rollout
Step-07: Traffic Splitting
- version-1: 33%
- version-2: 33%
- version-3: 34%
# Access Application on Browser using Revision URL
https://myservice1-czbx2i66ca-uc.a.run.app/
Observation: Traffic splits between 3 versions
Step-08: Verify Additional Tabs
- Verify Logs Tabs
- Verify Metrics Tabs
- Verify Security Tabs
Google Cloud Run Services – Autoscaling
Fast request-based auto scaling
- Minimum Instances: starts from zero, Set to 1 to reduce cold starts
- Maximum Instances: scale out to 1000 instances and more with a request to increase quota
Scale to zero and minimum instances
- When minimum instances set to zero and no requests then active instances will be zero
- New instance created as soon as the request comes in
- Negatively impacts the response times for the first request
Step-09: Cloud Run Autoscaling
Cloud Run Autoscaling
- Minimum number of instances
- Maximum number of instances
- Cold Starts
You can control autoscaling behavior with min/max instance limits, while cold starts occur when new instances spin up from zero.
Google Cloud Run Services – Access Modes
Ingress Control
- Public Service
Allow direct access from internet
- Private Service
Allow traffic from VPC
Allow traffic from external Application Load Balancers
Authentication
- Un-authenticated Access (Public API or Website)
- Authenticated Access using Cloud Identity-Aware Proxy (Secure access via web or mobile clients)
Step-10: gcloud: Create Google Cloud Run Service
# gcloud Project Settings
gcloud config list
PROJECT_ID=[YOUR-PROJECT-ID]
PROJECT_ID=gcpdemos
REGION=us-central1
gcloud config set core/project $PROJECT_ID
gcloud config set run/region $REGION
gcloud config list
# Help
gcloud run services --help
gcloud run deploy --help
# List Cloud Run Services
gcloud run services list
# Create Google Cloud Run Service
gcloud run deploy myservice102
--image=stacksimplify/google-cloud-run:v1
--allow-unauthenticated
--port=80
# List Cloud Run Services
gcloud run services list
# Describe Cloud Run Service
gcloud run services describe myservice102
Step-11: gcloud: List and Describe Revisions
# Help
gcloud run revisions --help
# List Revisions
gcloud run revisions list
# Describe Revision
gcloud run revisions describe
gcloud run revisions describe myservice102-00001-2rk
Step-12: gcloud: Update Application
# Update Application
gcloud run services update
gcloud run services update --help
# Update Application
gcloud run services update myservice102 --image=stacksimplify/google-cloud-run:v2
# List Revisions
gcloud run revisions list
# Describe Revision
gcloud run revisions describe
gcloud run revisions describe myservice102-00001-2rk
Step-13: gcloud: Update Traffic
# Help
gcloud run services update-traffic --help
# List Revisions
gcloud run revisions list
# Set Tags (Add Revision URLs)
gcloud run services update-traffic myservice102
--set-tags=myappv1=myservice102-00001-2rk,myappv2=myservice102-00002-xgl
# Update Traffic - V1-50%, V2-50%
gcloud run services update-traffic myservice102
--to-revisions=myservice102-00001-2rk=50,myservice102-00002-xgl=50
## 1. You can also refer to the current or future LATEST revision in --to-revisions by the string "LATEST".
## 2. To set 10% of traffic to always float to the latest revision:
gcloud run services update-traffic myservice102
--to-revisions=myservice102-00001-2rk=100,myservice102-00002-xgl=0
gcloud run services update-traffic myservice102 --to-revisions=LATEST=10
# To assign 100% of traffic to the current or future LATEST revision run
gcloud run services update-traffic myservice102 --to-latest
Step-14: gcloud: Delete Cloud Run Service
# List Cloudn Run Services
gcloud run services list
# Delete Cloud Run Service
gcloud run services delete myservice102
gcloud run services delete myservice1
