EDB Agent Governance is a standalone application that provides observability and oversight for AI agents operating on Postgres databases. It lets database administrators, security teams, and AI operations engineers monitor, audit, and reason about how AI agents interact with enterprise data.
The application connects to one or more upstream log sources — an EDB Hybrid Manager (HM) deployment or a standalone Grafana Loki endpoint — to retrieve the Postgres query logs that AI agents generate through the Airman Model Context Protocol (MCP) server. It presents those logs as structured sessions (logical groupings of SQL statements executed within a single AI agent interaction) with tools for filtering, inspecting, and auditing agent behavior. Because it reads from any qualifying log source, it isn't limited to HM-managed databases.
Exploring capabilities
- Reconstruct agent sessions — See every SQL statement an AI agent executed, with full metadata such as duration, target database, Postgres role, and error severity.
- Audit across multiple data sources — Connect to several HM or Loki sources at once and audit clusters across all of them from one place.
- Read a flat audit trail — Scan every statement in a session as a single filterable table, optimized for review.
- Rely on cached, incremental sync — The application incrementally syncs and caches session data, so audit views load quickly without requerying the full log history on every visit.
Getting started
- Key concepts — The key terms used throughout the application: sessions, steps, purpose, instances, and how Airman MCP tags activity.
- Architecture and data flow — How agent activity travels from Postgres logs into the application.
- Configuring — For administrators configuring the application after deployment.
- Connecting data sources — Register an HM or Loki instance so its agent sessions appear.
- Configuring Airman MCP — Set up Airman MCP instances with purpose labels, access modes, and PostgreSQL roles so governance boundaries are enforced.
- Browsing sessions — Find and filter the AI agent sessions detected for a cluster.
- Inspecting a session — Inspect a single session step by step and read its audit log.
- Securing access and handling credentials — Authentication, credential storage, and how your data is treated.
Using Agent Governance with Hybrid Manager
If your Postgres clusters are managed by HM, see Governance in Hybrid Manager for the HM-specific integration details — Machine user API keys, HM's Loki pipeline, and project and cluster discovery.
Key concepts
Key terms used throughout the AI governance audit log viewer — sessions, steps, purpose, instances, and how Airman MCP tags agent activity.
Architecture
How AI agent activity travels from Postgres logs through Loki and the viewer's backend into the audit log viewer, and how background sync keeps it fast.
Configuring
For administrators configuring EDB Agent Governance — managing identity providers and securing your deployment.
Connecting data sources
Register an EDB Hybrid Manager or standalone Loki instance so its AI agent sessions appear in the audit log viewer.
Configuring Airman MCP
Set up Airman MCP instances with purpose labels and access modes, and configure Postgres roles so governance boundaries are enforced at the database execution boundary.
Browsing sessions
Find and filter the AI agent sessions detected for a Postgres cluster in the audit log viewer.
Inspecting a session
Inspect a single AI agent session step by step and read its full audit log in the governance audit log viewer.
Securing access and handling credentials
How the AI governance audit log viewer authenticates users, stores upstream credentials, and treats your data.