Virtana Introduces Agentic Observability Enterprise Management for Enhanced IT Incident Response
Information technology observability firm Virtana has today unveiled its latest feature, Agentic Observability Enterprise Management, designed to allow its artificial intelligence agents to collaborate with agents from other vendors during IT incidents.
This new capability addresses a critical gap identified in the adoption of AI agents within enterprises. Traditional enterprise agents are often integrated into specific vendor products, such as IT service management systems. As a result, when failures occur across multiple vendor systems, these agents are unable to obtain a comprehensive view of the outage, leaving engineers to manually compile data from different sources.
With the introduction of this feature, a Virtana agent that detects an issue can now invite external vendor agents to join the investigation. Agents that comply with the Model Context Protocol or the open Agent2Agent protocol can participate, including those already bundled within the software that a company already uses.
Since February, external agents have had the capability to query Virtana’s monitoring data, following the launch of the System-Aware MCP Server. This server provides access to over 25,000 real-time metrics and enables invited agents to view how applications rely on their underlying infrastructure, thanks to the System Dependency Graph introduced concurrently. A month later, Virtana expanded its monitoring capabilities to include AI workloads.
In a practical scenario described by Virtana, a third-party agent may identify a slowdown in an application and solicit the assistance of Virtana’s root cause analysis agent via Agent2Agent. The investigation may reveal a database bottleneck arising from a recent software update. Follow-up queries can be directed back to the root cause agent as additional specialized agents are brought in to assist with the resolution. Virtana’s system then confirms that the application has returned to its normal operating state.
Amitkumar Rathi, Chief Product Officer at Virtana, noted that while a traditional IT war room may take hours to assemble and execute, a team of agents can complete the same processes “in milliseconds.”
This level of autonomy is limited to incident types predefined by customers. For businesses that prefer a more cautious approach, engineers can execute the recommended fixes from agents or require approval for each agent action before implementation. Specific policies also dictate which external agents are permitted in an investigation and the number of tokens they can use during their participation. Each action taken by the agents is logged for auditing purposes.
Virtana, previously known as Virtual Instruments until a rebranding in 2019, is supported by venture capital and was originally focused on providing tools for monitoring IT infrastructure performance. In 2022, the company secured $73 million in funding from investors including Atalaya Capital Management LP and Benhamou Global Ventures to advance its hybrid cloud observability initiatives.