Talk Nerdy to Me
Back to Blog
devopsawsaiopsreinvent-2025incident-responsesre

AWS DevOps Agent — The Autonomous On-Call Engineer

By Mathieu Kessler10 min read

AWS announced something at re:Invent 2025 that every DevOps engineer needs to know about: the AWS DevOps Agent (now in preview).

This isn't just another alerting tool. It's an autonomous on-call engineer that investigates incidents, identifies root causes, and coordinates response — all while you sleep.

86%

Root Cause ID Rate

Internal Amazon testing

Hours→Min

MTTR Reduction

Time to resolution

24/7

Autonomous Triage

Always investigating

What It Does

AWS DevOps Agent acts as an autonomous on-call engineer. It doesn't just alert you — it actively investigates incidents and coordinates response.

The moment an alert fires — whether at 2 AM or during peak hours — the agent begins investigating. It correlates data across your entire operational toolchain, from metrics and logs to recent code deployments.

AWS describes it as providing "always-on incident triage, guided resolution, and recommendations for how to continuously improve the reliability and performance of your applications across AWS, multicloud, and hybrid environments."

How It Works

The agent has deep knowledge of your applications and the relationships between components. It analyzes data across multiple sources simultaneously:

Observability Tools

  • • Amazon CloudWatch (metrics & logs)
  • • AWS X-Ray (traces)
  • • Datadog
  • • Dynatrace
  • • New Relic
  • • Splunk

Code & CI/CD

  • • GitHub repositories
  • • GitHub Actions
  • • GitLab repositories
  • • GitLab workflows
  • • Recent commits & PRs
  • • Deployment history

Incident Management

  • • ServiceNow (built-in)
  • • PagerDuty (via webhooks)
  • • Slack integration
  • • Custom ticketing systems

Extensibility (MCP)

  • • Bring Your Own MCP servers
  • • Grafana integration
  • • Prometheus integration
  • • Custom internal tools
  • • Open source solutions

The Investigation Process

1

Alert triggers — Agent begins investigating immediately, 24/7

2

Correlates data — Metrics, logs, traces, and recent code changes analyzed together

3

Identifies root cause — 86% accuracy rate based on internal Amazon testing

4

Provides mitigation plan — Engineers can implement recommended fixes

5

Interactive chat — Engineers can query the agent for more context at any time

Why This Matters

Traditional AIOps tools give you smarter alerts. AWS DevOps Agent goes further — it's an AI teammate that can triage, investigate, and act.

For on-call engineers, this means:

Faster MTTR

Mean time to resolution drops from hours to minutes. The agent starts investigating the moment an alert fires.

Fewer False Positives

By correlating data across multiple sources, the agent filters noise and surfaces real issues.

Context-Rich Handoffs

When humans need to intervene, they get full context of what the agent discovered, not just an alert.

Real-World Results

Commonwealth Bank of Australia

The bank's Cloud Foundations group manages over 1,700 AWS accounts and provides centralized cloud operation services for thousands of engineers.

They replicated a complex network and identity management issue to test the agent. These types of issues typically take a seasoned DevOps engineer hours to identify. The agent found the root cause in under 15 minutes.

Western Governors University

"At WGU about 200,000 students rely on 24/7 online learning, making system reliability critical to their success. To better serve our students, we implemented AWS DevOps Agent integrated with Dynatrace in our production environment, and the initial results are significant."

Internal Amazon

Within Amazon, AWS DevOps Agent has handled thousands of escalations, with an estimated root cause identification rate of over 86%.

The Vision: Autonomous Ops

AWS describes the end goal as reaching "a nirvana stage of autonomous ops":

"An alert goes off, but the agent has already figured out what the root cause is and has taken the corrective remediation action. You wake up in the morning and it just gives you a report of everything that happened last night and all the changes it made and all the fixes it made."

The Bigger Picture

AWS is betting big on autonomous agents. DevOps Agent joins Kiro (autonomous coding) and Nova Act (browser automation) as part of their "frontier agents" strategy.

The pattern is clear: AI is moving from assistants to actors.

We're not just getting help writing code anymore. We're getting help running production systems.

AWS Support Integration

Customers previewing AWS DevOps Agent can engage with AWS Support with one-click when needed, giving AWS experts immediate context of the situation for faster resolution.

Preview Availability & Limits

AWS DevOps Agent is available now in preview. During the preview period, AWS is not charging for the agent, but with monthly limits:

  • •10 Agent Spaces per account
  • •20 DevOps Agent incident response hours
  • •10 DevOps Agent incident prevention hours
  • •1,000 chat messages per month

The Question

Ready to hand over the pager to an AI agent, or keeping humans in the loop?

The technology is here. The question is whether your organization is ready to trust it.

Sources & Further Reading

Back to Blog