AWS ships DevOps Agent
AWS formally launched a DevOps Agent that supports incident response across Azure and on-prem environments, extending its tooling beyond AWS and signalling a push to be central to multi-cloud operations. The product aims to let teams run unified incident workflows regardless of where workloads live, addressing common cross‑cloud response gaps. (x.com)
When a website breaks at 2 a.m., the hard part is usually not fixing one server. The hard part is that the clues are scattered across alarms, chat logs, code changes, dashboards, and ticket systems, and many companies now spread those clues across Amazon Web Services, Microsoft Azure, and their own data centers. (aws.amazon.com) Amazon Web Services says its new DevOps Agent is built to do that hunting automatically. The service reached general availability on March 31, 2026, after a public preview announced on December 2, 2025. (aws.amazon.com 1) (aws.amazon.com 2) A DevOps tool is the software equivalent of a control room operator: it watches systems, follows runbooks, and helps engineers respond when something fails. Amazon Web Services describes DevOps Agent as an “always-available operations teammate” that investigates incidents, suggests fixes, and looks for ways to prevent the same outage from happening again. (aws.amazon.com 1) (aws.amazon.com 2) The new part is where it can look. In general availability, Amazon Web Services added the ability to investigate applications running in Microsoft Azure and in on-premises environments, instead of limiting the agent to Amazon’s own cloud. (aws.amazon.com) That Azure connection is not just a logo on a slide. Amazon’s documentation says the agent can use Azure Resource Graph to inspect Azure resources such as virtual machines, Azure Kubernetes Service clusters, databases, and network components during an investigation. (docs.aws.amazon.com 1) (docs.aws.amazon.com 2) It can also plug into Azure DevOps, which is Microsoft’s code and pipeline service. That lets the agent compare an outage with recent repository changes and pipeline runs, so it can test the oldest question in operations: did a deployment just break this? (docs.aws.amazon.com) Amazon is also wiring the agent into the systems companies already use to open incidents. Amazon says ServiceNow tickets can automatically trigger an investigation, and webhooks can start the same process from PagerDuty tickets or Grafana alarms. (aws.amazon.com) (docs.aws.amazon.com) Inside Amazon’s product, the organizing unit is called an Agent Space. That web app is where operators can watch investigations, ask questions in plain language, browse application topology across accounts, and review prevention ideas for code, observability, pipelines, and infrastructure. (docs.aws.amazon.com) Amazon is selling speed as much as automation. In its launch materials, the company says preview customers reported up to 75 percent lower mean time to resolution, and Western Governors University said incident resolution dropped from hours to minutes. (aws.amazon.com) The bigger move is strategic. For years, Amazon Web Services built tools that worked best when a company kept everything inside Amazon, but this launch openly targets the messier reality where one company might run apps in Amazon Web Services, databases in Microsoft Azure, and older systems in its own server rooms. (aws.amazon.com) (aws.amazon.com) If Amazon can become the place where engineers investigate outages no matter where the outage started, it gets a seat above the infrastructure layer. That is a different ambition from renting servers: it is trying to become the operating console for multi-cloud operations. (aws.amazon.com) (aws.amazon.com)
