Infrastructure-aware agents via MCP and knowledge graph

As teams scale, static runbooks and brittle scripts don’t cut it anymore. That’s why we’ve been working on a better way for Aptible AI to represent your system dynamically and let agents interact with it like an experienced engineer would.

MCP server progress

• We introduced a prototype of our Model Context Protocol (MCP) server, which allows agents to traverse the knowledge graph dynamically and take action.

• Example capabilities include:

  • Mapping system topology to understand architecture

  • Fetching details on specific services using tools like dig or curl

  • Handling ambiguous requests through intelligent trial-and-error

Diagnostic tooling updates

• Diagnostic Dashboards now respect local timezones, making it easier for global teams to interpret timelines.

• Users can now add free-form symptom descriptions to provide context for AI-generated summaries.

This work paves the way for truly interactive diagnostics where Aptible AI doesn’t just observe, but investigates your system as it changes.