OpsRabbit Trust Center

Data Flow and Data Handling

How information moves through a customer-hosted OpsRabbit deployment.

PublicVersion 1.0September 2026

Overview

This document follows information from an authorized customer source through retrieval, model inference, output and audit logging. It is designed to help security and privacy reviewers identify trust boundaries and control owners.

Actual systems, regions, providers, retention and integrations are confirmed in the deployment record.

Primary flow

  1. Authorized operational systems expose signals through customer-approved integrations.
  2. OpsRabbit retrieves only information permitted by configured identities and scopes.
  3. The product assembles relevant evidence and trusted instructions.
  4. Required context is sent to the customer-approved model endpoint for inference.
  5. OpsRabbit presents evidence-backed analysis and records relevant audit events.
  6. Any authorized action follows applicable permission and approval controls.

Data categories

  • Operational alerts, logs, metrics, tickets and service context
  • User, role and approval metadata
  • Model context, output and supporting evidence
  • Tool-call, policy-decision, error and security events

Boundary controls

Customer operational data remains in the approved customer-hosted and provider boundaries. Egress, storage, retention, support access and backups depend on the agreed architecture. Secrets should be supplied through managed connections and excluded from model context.

Data inventory

StageTypical dataPrimary control
Source accessAlerts, metrics, logs, traces, tickets, topology and service metadataCustomer-approved identity and scope
Retrieval and correlationTask-relevant evidence and relationshipsMinimization, authorization and isolation
Model contextSelected evidence, instructions and task metadataTrusted-instruction separation, secret exclusion and approved endpoint
OutputAnalysis, evidence citations and recommendationsHuman validation and action controls
AuditActor, tool, model, policy, approval and result metadataRestricted access, integrity and configured retention

Support and administrative access

OpsRabbit provider personnel do not receive standing access to customer operational data solely because OpsRabbit is deployed. Any support access must be authorized, time-bounded, least-privilege and logged. The customer should remove access at the end of the activity. Customer data should not be transferred into general support channels where a secure alternative is available.

Cross-border and provider flows

The deployment record identifies where OpsRabbit components, customer data and model inference operate. Customer-selected providers may create separate cross-border transfers and retention obligations. OpsRabbit provider does not change the approved provider or processing location without the governance and notice required by the agreement.

Deployment exit

At the end of the service or evaluation, the parties disable temporary accounts and integrations, revoke credentials, return or delete OpsRabbit provider-controlled customer data, preserve only required security or contractual evidence, and confirm responsibility for customer-hosted databases, logs and backups.