Get Your CertDatabricks Certified Machine Learning ProfessionalPlayCompete
‹ All Databricks Certified Machine Learning Professional questions

Databricks ML Professional Practice Question 10: Canary Traffic Split

Databricks Certified Machine Learning Professional · Test 1 · question 10 of 10

Deployment Strategieseasy

A Model Serving endpoint named recs sends 100% of its traffic to version 12 of a recommendation model. The team wants a canary rollout of version 13 with these requirements:

- 10% of live requests go to version 13. - Client applications keep calling the same endpoint URL, with no code changes. - QA can send test requests only to version 13. - Rollback means just changing the endpoint configuration.

How should the team do this?

Select one answer
Show answer and explanation

Answer:

  • D. Add version 13 as a second served model on recs, with routes of 90/10, and have QA call that served model's own invocations path

One Model Serving endpoint can host several served models of the same type. Its traffic_config routes split incoming requests between them by percentage, for example 90% to the current version and 10% to the challenger. Clients keep calling /serving-endpoints/recs/invocations, and you can change the split, or send everything back to version 12, with a config update through the UI or PUT /api/2.0/serving-endpoints/{name}/config. QA can skip the routing and call one model directly at /serving-endpoints/recs/served-models/<served-model-name>/invocations.

Option A moves the traffic split into the client code, which the requirements forbid.

Option B confuses registry metadata with serving configuration. Aliases do not create traffic routes on an endpoint. Routes are defined in the endpoint's config.

Option C is a big-bang cutover, not a canary. Every user is exposed to version 13 at once, and rolling back means another full redeployment.

Reference: Serve multiple models to a model serving endpoint

Practice Test 1 ›

Study guides for Databricks Certified Machine Learning Professional

Best MLOps certifications 2026: which test real MLOpsThe best MLOps certifications in 2026 from Microsoft, Databricks, Google and AWS, ranked by how much each exam tests MLOps, with cost and order.
Best AI certifications 2026: AWS, Google, Microsoft, ClaudeThe best AI certifications in 2026 from AWS, Google, Microsoft, Claude and Databricks, by level, with price, status and which one to take first.