July 2026 Summaries

Guillaume Da Silva, a Staff Solutions Architect at Qovery, discusses how the company's MCP server and AI skills enhance troubleshooting of customer production incidents by transforming vague issues into detailed, evidence-backed diagnoses. By leveraging the MCP server, agents gain access to real infrastructure data, enabling them to systematically gather and analyze information about organizations, environments, and services without making autonomous changes. The Qovery Skills provide structured diagnostic processes that help agents understand and resolve issues, such as container memory limits being exceeded, by correlating Kubernetes events, deployment history, and application logs. This approach allows engineers to maintain control over production environments while ensuring that every investigation starts from a consistent baseline, ultimately turning support tickets into actionable insights and facilitating infrastructure optimization through informed recommendations.

AI agents are increasingly integrated into organizational structures, prompting the need for stringent governance measures, which Qovery facilitates by enabling AI workflows to operate within a company's own infrastructure. This ensures data control, limited access, and auditability, with governance enforced at the network level through a custom layer 7 proxy that restricts and audits all outbound communications. This setup allows for the orchestration of valuable workflows like incident response, cost analysis, and feature bootstrapping, enhancing operational efficiency and compliance while maintaining data security. By focusing on network-level governance, Qovery empowers organizations to harness AI's potential without compromising control over their data and infrastructure, positioning itself as a key platform for managing AI workflows in production environments.

Qovery Blueprints provide a solution for managing infrastructure provisioning by AI agents and developers, ensuring control over deployments with built-in access and cost rules. These Blueprints are reusable and composable templates for infrastructure blocks like MySQL, PostgreSQL, S3, and Redis, allowing for self-service without compromising governance. They address the speed versus control dilemma faced by platform teams by enabling both agents and developers to deploy infrastructure under the same rules, thus removing the need for manual intervention or extensive security reviews. By encoding governance within the platform, Qovery Blueprints streamline the provisioning process, support cloud portability, and reduce the risk of unmanaged infrastructure, as updates to Blueprints automatically propagate to all instances.