Integrations

MCP integration

Senal Ops exposes a hosted Model Context Protocol (MCP) server so an AI assistant can read Recover backup status and monitor health, and trigger a subset of actions, using the same API key as telemetry ingestion.

Audience: Developers, AI engineers, platform teams

Connecting a client

  • Create an API key from the API keys page and copy the raw value; it is shown once.
  • Add the MCP server URL, https://senalops.com/api/mcp, to your MCP client configuration.
  • Send the API key as a Bearer token in the Authorization header, the same way telemetry ingestion authenticates.
  • No separate MCP-specific credential is required; revoking the API key also disconnects any MCP client using it.

Available tools

  • list_connections: lists protected database connections and their status. Requires a plan with Recover.
  • list_backups: lists backup job history and validation status, optionally filtered by connection. Requires a plan with Recover.
  • trigger_backup: triggers a manual backup for a protected connection. Requires a plan with Recover.
  • list_monitors: lists monitors with current up, down, or degraded status. Available on every plan within its monitor limit.
  • list_incidents: lists incidents for a specific monitor.
  • create_monitor: creates an HTTP, TCP, ping, heartbeat, or AI-agent monitor, enforcing the same plan limits and target-safety checks as the dashboard.

Scope and safety

  • Every tool call is scoped to the organization that owns the API key; no tool accepts an organization ID as an argument.
  • Restoring a backup and pausing or resuming a monitor are dashboard-only actions and are not exposed through MCP.
  • trigger_backup and create_monitor calls are written to the organization audit log, distinguishable from dashboard-triggered actions.
  • MCP requests share the same rate limits and plan entitlement checks as the dashboard and ingestion endpoints.

Troubleshooting

  • If a Recover tool returns a plan error, confirm the organization's plan includes Recover.
  • If a request is rejected with an authentication error, confirm the Authorization header uses a valid, non-revoked API key.
  • If create_monitor rejects a target, confirm the URL or host is publicly reachable and not a private or internal address.
  • Review the audit log to confirm whether an action was triggered through MCP or the dashboard.

Related documentation