Question

Your team deploys a monolithic web application to a fleet of 12 on-premises IIS servers behind a load balancer. You need an Azure Pipelines mechanism that installs an agent on each server, targets the machines by tags (for example 'web' and 'db'), and rolls out the deployment to a percentage of the servers at a time so the application stays available throughout. Which Azure Pipelines feature should you use to target and orchestrate these physical servers?

Accepted Answer

Create a deployment group, register each server as a target with tags, and reference it in a deployment job that uses the 'runOnce' or 'rolling' strategy with a batch percentage.

Answer

Create an environment with a Kubernetes resource and reference it in a deployment job so pods roll out gradually.

Answer

Create a self-hosted agent pool and set the parallel job count to 12 so all servers deploy simultaneously.

Answer

Use a Microsoft-hosted agent with a matrix strategy that iterates over server hostnames using WinRM.