Amazon OpenSearch Service has introduced MCP Apps, a set of Model Context Protocol-based observability tools that let compatible AI coding environments query and visualize logs, traces, metrics, alerts, Prometheus data, service maps, and stack health during incident investigation.
Amazon Web Services has launched MCP Apps for Amazon OpenSearch Service, adding Model Context Protocol-based observability workflows for compatible AI development environments.
In a June 2026 “What’s New” post, AWS said Amazon OpenSearch Service now includes MCP Apps for “agentic observability.” According to AWS, the feature is intended to let developers and operators investigate incidents from inside a conversational AI development environment rather than switching between separate monitoring screens and query tools.
AWS says the apps can work with observability data in OpenSearch, including logs, traces, metrics, alerts, Prometheus data, and interactive visualizations. The company describes the experience as one where a user can ask questions about an operational issue and receive both text-based answers and visual context inside the assistant conversation.
The Amazon Web Services developer guide describes MCP Apps as part of an OpenSearch observability workflow built around the Model Context Protocol, or MCP. MCP is used here to expose OpenSearch capabilities as tools that an AI assistant or agent framework can call during a troubleshooting session.
AWS documentation says the OpenSearch MCP server is an open-source MCP implementation that exposes OpenSearch APIs as tools for AI assistants and agent frameworks. The same documentation says it can be used with managed Amazon OpenSearch Service domains and OpenSearch Serverless collections.
The observability-specific guide lists app categories including logs, traces, metrics, service maps, and stack health. It also describes “dual text-and-visual responses,” meaning the assistant can return a natural-language explanation alongside visual output relevant to the investigation.
The practical goal is to reduce the friction of moving between monitoring systems, query languages, dashboards, and code editors during an outage or performance investigation. AWS’s announcement frames MCP Apps as a way for compatible agentic IDEs to investigate incidents using multiple telemetry sources within the same conversation.
That does not mean the AI system replaces operational judgment. The source documents describe tool access, workflows, prerequisites, and supported categories; they do not claim autonomous remediation or guaranteed diagnosis. The more concrete change is that OpenSearch observability data can be made available to an AI assistant through documented MCP tooling.
For teams already using Amazon OpenSearch Service for operational analytics, the feature could make incident exploration more interactive. A developer might ask about recent error spikes, related traces, affected services, or metric changes, while the assistant calls OpenSearch-backed tools to retrieve supporting information. AWS’s documentation indicates that the workflow can include both analytical answers and visualizations, which may help users validate findings during a live investigation.
AWS’s developer documentation says users must set up the OpenSearch observability MCP Apps workflow and use a compatible IDE or assistant environment. The docs cover prerequisites, configuration steps, and the available app categories.
AWS also documents the OpenSearch MCP server separately, positioning it as the component that exposes OpenSearch APIs to MCP-compatible assistants and frameworks. Organizations evaluating the feature should review those setup requirements, supported environments, and security controls before connecting production observability data to an AI assistant workflow.
The launch follows a broader industry move toward connecting AI coding assistants and operational tools through MCP. In this case, AWS is applying the protocol to observability data held in OpenSearch Service and OpenSearch Serverless.
The announcement is notable because it brings structured logs, traces, metrics, alerts, Prometheus data, service maps, and health views into an AI-assisted troubleshooting flow. Based on AWS’s announcement and documentation, the feature is best understood as a new interface for querying and visualizing existing OpenSearch observability data, rather than a standalone monitoring product.
Amazon Web Services has launched MCP Apps for Amazon OpenSearch Service, adding Model Context Protocol based observability workflows for compatible AI development environments.
AWS says the apps can work with observability data in OpenSearch, including logs, traces, metrics, alerts, Prometheus data, and interactive visualizations.
The company describes the experience as one where a user can ask questions about an operational issue and receive both text based answers and visual context inside the assistant conversation.
Continue reading
