Azure Virtual Machines carry the workloads your business depends on, but Azure decides how to manage them. Planned maintenance, host updates, and hardware retirement arrive on a schedule set by the platform, not by your change calendar. Each one is a small test of whether your application recovers on its own.

Install the Gremlin Agent on your VMs and you can run that test whenever you choose, with Gremlin recording the result and recommending remediations. With Gremlin, you can:

• Survive the maintenance you don't schedule. Shutdown experiments take a VM offline on your terms. You'll learn how quickly the load balancer stops sending traffic, whether your application restarts cleanly, and how much state disappears with the instance.
• Size your VMs against real pressure. CPU, memory, and disk experiments push a VM to its limits while you watch response times. This is where teams often discover their scaling thresholds sit well above the point where customers start noticing.
• Prove zone redundancy holds. Blackhole experiments isolate an availability zone so you can confirm the remaining zones absorb the traffic. Most teams treat zone redundancy like backups: spend the time and money to set it up, but don't validate that it works until it's too late.

Azure decides when your VMs move. You decide whether your application is ready when they do.