Data Science Model Deployments and Cloud Computing on GCP - Lab - Traffic Splitting in App Engine

The video tutorial explains how to split traffic between different versions of an application, a common scenario for testing new versions while maintaining stability. It covers two methods: splitting by IP address and splitting randomly. The tutorial provides step-by-step instructions for implementing traffic splitting using gcloud commands, testing the setup, and verifying the results. The video concludes with a preview of the next tutorial, which will focus on deploying an app that reads data from BigQuery using Flask API.

Why is it important to split traffic between different versions of an application?

• To ensure all users see the same version

• To test new versions while keeping the stable version active

• To increase application speed

• To reduce server load

What command is used to set traffic splits in gcloud?

• gcloud app services set traffic

• gcloud app deploy

• gcloud app update

• gcloud app rollback

What percentage of traffic is allocated to each version in the example provided?

• 70% to V1 and 30% to V2

• 60% to V1 and 40% to V2

• 100% to V1 and 0% to V2

• 50% to V1 and 50% to V2

What is the default method of traffic splitting if no specific method is mentioned?

• By session duration

• By user ID

• By IP address

• Randomly

How can you verify if the random traffic splitting is working correctly?

• By deploying a new version

• By refreshing the browser multiple times

• By using a different IP address or device

• By checking the server logs