AWS Cloud Operations Blog

This Month in AWS DevOps Agent: November 2026

Introduction

November brought AWS DevOps Agent to AWS GovCloud (US-West) and added AWS Organizations support, so teams can connect an entire organizational unit to an Agent Space instead of one account at a time. Investigations now complete faster and show a real-time summary as they run. Replay and automatic Reviews give teams a way to measure investigation quality and test configuration changes against past incidents. Memory management expanded with custom memory imports, search by meaning, a knowledge graph, and memory dreaming, which keeps managed memories accurate over time. Chat extended to Microsoft Teams, Slack code channels, and Model Context Protocol (MCP) clients, and sandboxes let the agent write and run code. Partner-managed connectors and the Skills Catalog make it easier to extend the agent.

Here’s what’s new across AWS DevOps Agent. For the full list of changes, see the AWS DevOps Agent release history page.

Investigate faster with real-time summaries and sandboxes

The clearest theme this month is shortening the time from alert to answer.

Investigations performed by AWS DevOps Agent now complete faster. Each investigation includes a real-time summary on its Summary tab that updates as the agent collects evidence, so initial findings, the current cause hypothesis, and next steps surface before the final root cause and mitigation plan. Customers who use the built-in Slack integration receive these summaries in the relevant Slack channel. To learn more, read the What’s New post.

AWS DevOps Agent Sandbox, now in preview, gives the agent an isolated Linux environment where it can write and run code during investigations and chat conversations. The agent can use code to analyze a large log file, calculate statistics across a set of metrics, or correlate data from multiple sources. Each sandbox is scoped to a single investigation or conversation and is discarded when it ends. AWS calls from the sandbox are limited to read-only operations, and by default the sandbox can reach only AWS service endpoints until you add hosts to its allowlist. With Sandbox enabled, your skills can include executable code, such as your team’s own diagnostic scripts. During preview, Sandbox is available in US East (N. Virginia), US West (Oregon), Asia Pacific (Tokyo), and Europe (Ireland). To get started, see Set up a Sandbox.

AWS DevOps Agent also now publishes nine investigation metrics to Amazon CloudWatch in the AWS/AIDevOps namespace, with no configuration required. The metrics count investigations by end status, count completed investigations that found a cause or have a mitigation plan, record the time to the first cause and first mitigation plan, and record the result of each Review. You can create CloudWatch alarms on these metrics. For the full list, see Vended Logs and Metrics.

Measure and improve investigation quality

Three launches help teams see how well AWS DevOps Agent performs and validate changes before relying on them.

ou can now replay a completed investigation after you change skills, instructions, or integrations, then compare the replay with the original by grader score, time to root cause, tools and skills used, and investigation timeline. Replays take no mutative actions and do not change the original. Because replays query your live environment, data available at the time of the original investigation might not be available to the replay. To learn more, see Replaying investigations.

A Review is a quality assessment of a completed investigation. It returns a helpful or not helpful verdict and a pass, fail, or not applicable result for five dimensions: scope alignment, evidence grounding, completeness, environment impact, and accuracy. AWS DevOps Agent now runs Reviews automatically on a sample of completed investigations, when a linked ServiceNow or PagerDuty ticket is resolved, 7 days after an investigation completes, or both. Automatic Reviews are on by default with a 10 percent sampling rate, which you can set from 1 to 100 percent on the Grader agent’s Configuration tab. To learn more, see Reviewing investigations.

The Incident RCA and Incident mitigation agents each have a new Dashboard tab. The Summary view shows metrics such as investigations, the percentage that found a cause, and the median time to first cause for a single time window. The Trend view plots these numbers over spans from 7 days to 6 months. To learn more, see Managed agents.

Code Change Risk Analysis, previously called release readiness review, is now in preview in all commercial AWS Regions where AWS DevOps Agent is available. It reviews code changes for production risks before deployment by checking each change against your standards, its dependencies across repositories, its access controls, and your production environments. Each review produces a report with a recommended action, risk findings with affected code locations, and recommendations. Reviews can run in Kiro and Claude Code, automatically on GitHub and GitLab pull requests and merge requests, and on demand through chat. You can set the review depth for each branch, choosing static code analysis or static analysis with verification testing in an environment that AWS DevOps Agent manages. To learn more, see Code Change Risk Analysis.

Give the agent better memory

Memory launches this month focus on bringing in your own knowledge and keeping what the agent learns accurate.

You can now add your own knowledge to a custom memory store in three ways. Upload Markdown files that are already in the memory format, import runbooks, postmortems, and other documents through chat, or keep a store in sync with a source by running a custom agent, such as the Confluence memory sync agent. When you import through chat, it proposes memories and a store for each, and saves nothing until you confirm. To learn more, see Custom memories.

Search on the Memories tab now matches the names and descriptions of stores and memories by meaning as well as by exact text, with text matches listed first. Agents use the same search to find memories during their work. To learn more, see Searching memories.

AWS DevOps Agent now links related memories into a knowledge graph, such as an alarm, the component that owns it, and a reflection on a past mistake with that alarm. Agents follow these links to find connected context, and you can browse them by choosing Relationships on the Memories tab. A new weekly task, memory dreaming, reviews how agents used each claim in your managed memories during recent investigations and custom agent runs. It then deletes claims that no longer hold, merges duplicates, and resolves contradictions. Every change records its reason and evidence and creates a new memory version you can review, and a directive can protect a claim or block a merge. To learn more, see Managed memories.

Extend the agent with partner connectors and published skills

Several launches reduce the setup work required to connect AWS DevOps Agent to the tools your team already uses.

Partner-managed connectors connect AWS DevOps Agent to an MCP server that a partner hosts. The connector defines the server’s endpoints and authentication methods, so you choose an endpoint and enter only your credentials. You can install the skills the partner recommends, and some connectors create an alert webhook that starts investigations from the partner’s alerts. The partner operates the MCP server, and requests to it can leave the AWS Region of your Agent Space. The Capability Providers page now lists all providers in one section that you can search and filter by category. To learn more, read the What’s New post and see Using partner-managed connectors.

For more visit, Add community skills and partner connectors to AWS DevOps Agent.

The Skills Catalog in the AWS DevOps Agent web app lists published skills, such as skills that investigate AWS Health events or check service quotas. You can read a skill’s files and add it to your Agent Space in one step, without a GitHub account or a zip file. Catalog skills do not update automatically, and you can sync them to the current version. To learn more, see Skills Catalog.

Amazon EC2 Image Builder now integrates with AWS DevOps Agent in the Image Builder console. You can troubleshoot up to 50 failed image pipelines at once and create resources by describing what you want. The agent reads build logs, workflow steps, components, and network configuration, then provides the root cause and the steps to fix it. To learn more, read the What’s New post.

Collaborate where your team already works

Chat launches bring AWS DevOps Agent into the channels and tools your team uses during an incident.

AWS DevOps Agent chat now extends to MCP clients such as Kiro and Claude Code, as well as Slack and Microsoft Teams channels. Directed chat actions support GitHub, GitLab, Azure DevOps, ServiceNow, and PagerDuty MCP tools, and chat can use a sandbox to run scripts. You can also rename chat conversations and archive them without deleting anything. To learn more, read the What’s New post.

Slack code channels give each investigation its own channel. Choose Open in code channel on an investigation card, and your team and the agent work on the investigation together without mentioning the agent in each message. In Microsoft Teams, users previously received only one-way notifications. Now you can mention AWS DevOps Agent in a private or shared channel to ask questions, explore your AWS resources, and steer investigations, and the agent replies in a thread. Before AWS DevOps Agent posts in a newly associated Slack or Teams channel, a channel member must authorize the association. To learn more, read the What’s New post and see Connecting Microsoft Teams.

Govern access across accounts and Regions

AWS DevOps Agent is now available in AWS GovCloud (US-West). An Agent Space in this Region can be associated only with AWS GovCloud (US) accounts, and some features work differently or are not available, including Sandbox and private connections. To learn more, read the What’s New post and see Supported Regions.

With AWS Organizations support, you can connect an organizational unit (OU) or your organization root to an Agent Space instead of connecting accounts one at a time. AWS DevOps Agent deploys and maintains an access role in each covered account through AWS CloudFormation StackSets, and covers accounts that later join the OU. Each Agent Space supports up to 20 associations and 3,000 covered accounts. Organization mode is not available in the AWS GovCloud (US) Regions. To learn more, see Connecting your AWS Organization.

Access Profiles link AWS IAM Identity Center users in your Agent Spaces to permission sets, so you can limit who can create or update skills and memories. When chat works with assets such as skills, memories, and custom agents, it now uses operator permissions. The content safety guardrail, powered by Amazon Bedrock Guardrails, screens prompts sent to your Agent Space for prompt injection and jailbreak attempts. It is on by default, and you can turn it on or off for each Agent Space. To learn more, see AWS DevOps Agent Security.

Conclusion

November’s launches expand where AWS DevOps Agent runs, how it learns, and how teams govern it. To get started, see the AWS DevOps Agent User Guide. For pricing, see the AWS DevOps Agent pricing page.

Heading to re:Invent? To see the session catalog for  AWS DevOps Agent, visit re:Invent event catalog. Or need help knowing what sessions to attend? Visit Operating in the agentic era: AIOps sessions at re:Invent 2026.

Nereida Woo

Nereida Woo

Nereida Woo is a Senior Cloud Ops Specialist Solutions Architect. She leads Agentic Operations with AWS DevOps Agent, working with customers, partners, and field teams to help organizations adopt AI-driven operations.

Matheus Arrais

Matheus Arrais

Matheus Arrais is the Worldwide Tech Lead for Cloud Operations at AWS, where he sets the global direction for an internal community of hundreds of AWS experts advancing the field of cloud operations spanning observability, AIOps, and the emerging AgenticOps paradigm. He partners closely with AWS service teams to design solutions at scale that help customers implement and support complex cloud infrastructure, bringing agentic capabilities to operations while building the operational foundations that AI itself depends on. Find him on LinkedIn (https://www.linkedin.com/in/matheusarrais/) or his personal website https://matheusarrais.com