What is the main benefit of managing Azure Machine Learning workspaces with Bicep?
Why D is the answer
Environments are repeatable and reviewed in Git
Infrastructure as code makes each environment reproducible, and changes go through review.
In detail
- Bicep makes environments repeatable.
- Changes are reviewed in Git.
- Reviewing infrastructure catches mistakes early.
Use case
An illustrative scenario, not a real incident.
A workspace change once broke production unnoticed. With Bicep, each change is reviewed in Git, and environments match.
See it in different ways
flowchart LR A["Bicep in Git"] --> B["GitHub Actions"] B --> C["Dev workspace"] B --> D["Test workspace"] B --> E["Prod workspace"]
sequenceDiagram participant D as Developer participant G as GitHub participant A as GitHub Actions participant Z as Azure D->>G: Merges a Bicep change G->>A: Starts the workflow A->>Z: Signs in with OpenID Connect A->>Z: Deploys to dev, test and prod
mindmap
root(("Bicep"))
n1["Repeatable"]
n2["Same result each time"]
n3["Reviewed"]
n4["Changes in Git"]Key point: Bicep gives repeatable, reviewed environments.
Why the other options are wrong
- A. Models no longer need any monitoring in production
- Monitoring is still needed.
- B. Notebooks run with no compute attached to them
- Notebooks still need compute.
- C. Training runs on cheaper hardware by default
- It has no effect on hardware price.