Revision management in GCP - Mini Project: Build & Apply

Revision Management with Google Cloud Run

📖 Scenario: You are managing a web application deployed on Google Cloud Run. You want to keep track of different versions (revisions) of your service so you can control traffic and roll back if needed.

🎯 Goal: Build a simple Google Cloud Run service configuration that creates a service with multiple revisions and manages traffic splitting between them.

📋 What You'll Learn

Create a Cloud Run service configuration with a specific service name

Define two revisions with different container images

Set traffic split between the two revisions

Ensure the configuration is valid and deployable

💡 Why This Matters

🌍 Real World

Managing revisions in Cloud Run helps deploy new versions safely and control traffic during updates.

💼 Career

Cloud engineers and DevOps professionals use revision management to maintain application availability and perform smooth rollouts.

1

Create initial Cloud Run service configuration

Create a YAML configuration file named service.yaml with a service named my-cloud-run-service and specify the API version serving.knative.dev/v1 and kind Service.

GCP

apiVersion: serving.knative.dev/v1
kind: Service
metadata:
  name: my-cloud-run-service
# Your code here

Hint

Start by defining the service metadata and the first revision container image.

2

Add a second revision with a new container image

Add a second revision by defining a new template under spec with the container image gcr.io/my-project/my-app:v2 and name it revision-v2 under metadata.

GCP

apiVersion: serving.knative.dev/v1
kind: Service
metadata:
  name: my-cloud-run-service
spec:
  template:
    spec:
      containers:
      - image: gcr.io/my-project/my-app:v1
# Your code here

Hint

Define the second revision template with its own metadata name and container image.

3

Configure traffic splitting between revisions

Add a traffic section under spec to split traffic 70% to the first revision named my-cloud-run-service-00001 and 30% to the second revision named revision-v2.

GCP

apiVersion: serving.knative.dev/v1
kind: Service
metadata:
  name: my-cloud-run-service
spec:
  traffic:
  # Your code here
  template:
    metadata:
      name: revision-v2
    spec:
      containers:
      - image: gcr.io/my-project/my-app:v2

Hint

Use the traffic field to assign percentages to each revision.

4

Complete the Cloud Run service configuration

Ensure the YAML file ends with the correct indentation and includes the containers section under the second revision template with the image gcr.io/my-project/my-app:v2.

GCP

apiVersion: serving.knative.dev/v1
kind: Service
metadata:
  name: my-cloud-run-service
spec:
  traffic:
  - revisionName: my-cloud-run-service-00001
    percent: 70
  - revisionName: revision-v2
    percent: 30
  template:
    metadata:
      name: revision-v2
    spec:
      containers:
      # Your code here

Hint

Double-check the indentation and container image under the second revision template.

Practice

1. What is the main purpose of revision management in Google Cloud services?

easy

A. To track and control different versions of your cloud services

B. To increase the storage capacity of your cloud database

C. To monitor network traffic between services

D. To automatically scale the number of virtual machines

Solution

1. Step 1: Understand revision management concept

   Revision management is about keeping track of different versions of your cloud services.

2. Step 2: Identify the main goal

   The main goal is to control versions for safe updates and rollbacks.

3. Final Answer:

   To track and control different versions of your cloud services -> Option A

4. Quick Check:

   Revision management = version control [OK]

Hint: Revision management means version control of services [OK]

Common Mistakes:

• Confusing revision management with scaling
• Thinking it manages storage or network
• Mixing it with monitoring tools

2. Which gcloud command is used to list all revisions of a Cloud Run service named my-service?

easy

A. gcloud run revisions list --service=my-service

B. gcloud run services list --revisions=my-service

C. gcloud revisions describe my-service

D. gcloud run deploy revisions --service=my-service

Solution

1. Step 1: Identify correct command structure

   The command to list revisions uses gcloud run revisions list with the service flag.

2. Step 2: Match service name parameter

   Using --service=my-service correctly specifies the service.

3. Final Answer:

   gcloud run revisions list --service=my-service -> Option A

4. Quick Check:

   List revisions = gcloud run revisions list --service [OK]

Hint: List revisions with 'gcloud run revisions list --service' [OK]

Common Mistakes:

• Using 'services list' instead of 'revisions list'
• Mixing 'describe' with listing
• Wrong command order or flags

3. Given the following commands, what will be the output of gcloud run revisions list --service=my-app after deploying twice?

gcloud run deploy my-app --image=image:v1 --no-traffic
gcloud run deploy my-app --image=image:v2 --no-traffic

medium

A. One revision with image:v2

B. No revisions because traffic is zero

C. Two revisions: one with image:v1 and one with image:v2

D. Error because --no-traffic disables revisions

Solution

1. Step 1: Understand deployment with --no-traffic

   Deploying with --no-traffic creates a new revision but does not route traffic to it.

2. Step 2: Analyze multiple deployments

   Deploying twice with different images creates two separate revisions.

3. Final Answer:

   Two revisions: one with image:v1 and one with image:v2 -> Option C

4. Quick Check:

   Each deploy creates a revision even with no traffic [OK]

Hint: Each deploy creates a revision, traffic doesn't affect revision count [OK]

Common Mistakes:

• Assuming no traffic means no revision
• Thinking only latest revision exists
• Confusing traffic routing with revision creation

4. You deployed a new revision but it is not receiving traffic. Which command fixes this by routing 100% traffic to the latest revision named my-app-00002?

medium

A. gcloud run revisions set-traffic my-app-00002 --percent=100

B. gcloud run services update-traffic my-app --to-revisions my-app-00002=100

C. gcloud run services update my-app --traffic my-app-00002=100

D. gcloud run services update-traffic my-app --to-revision my-app-00002=100

Solution

1. Step 1: Identify correct command to update traffic

   The command to update traffic routing is gcloud run services update-traffic.

2. Step 2: Use correct syntax for traffic split

   The syntax --to-revisions my-app-00002=100 routes 100% traffic to that revision.

3. Final Answer:

   gcloud run services update-traffic my-app --to-revisions my-app-00002=100 -> Option B

4. Quick Check:

   Update traffic with 'services update-traffic' and '--to-revisions' [OK]

Hint: Use 'gcloud run services update-traffic' with '--to-revisions' [OK]

Common Mistakes:

• Using 'update' instead of 'update-traffic'
• Wrong flag names like '--percent'
• Trying to set traffic on revision directly

5. You want to roll back your Cloud Run service web-app to a previous revision web-app-00003 safely without downtime. Which sequence of commands achieves this?

hard

A. gcloud run deploy web-app --image=web-app-00003

B. gcloud run services update web-app --revision web-app-00003

C. gcloud run revisions delete web-app-00004 && gcloud run services update-traffic web-app --to-revisions web-app-00003=100

D. gcloud run services update-traffic web-app --to-revisions web-app-00003=100

Solution

1. Step 1: Understand rollback via traffic routing

   Rolling back means routing traffic back to an existing previous revision.

2. Step 2: Use traffic update command

   Use gcloud run services update-traffic to send 100% traffic to the old revision.

3. Step 3: Avoid deleting revisions or redeploying

   Deleting or redeploying is unnecessary and may cause downtime.

4. Final Answer:

   gcloud run services update-traffic web-app --to-revisions web-app-00003=100 -> Option D

5. Quick Check:

   Rollback by routing traffic to old revision [OK]

Hint: Rollback by updating traffic to old revision, no redeploy needed [OK]

Common Mistakes:

• Trying to redeploy old revision as image
• Deleting newer revisions unnecessarily
• Using wrong commands to update revision