Deployment Architecture
Enterprise security teams need deep visibility into AI agent behavior, but shipping raw execution logs to a third-party cloud is often a non-starter. Compliance requirements, data sovereignty regulations, and the sensitivity of agent execution traces make this a real concern.
Arrakis addresses this with a hybrid deployment model that keeps sensitive data processing inside the customer’s environment while delivering the full platform experience as SaaS. A lightweight, containerized on-premise component handles data ingestion and detection locally. Raw execution data never leaves the customer’s security boundary. Only processed results (alerts, risk scores, posture metrics) are sent to the SaaS platform for dashboarding, policy management, reporting, and integrations.
Deployment Options
| Model | Data Ingestion | Detection | Platform | Best For |
|---|---|---|---|---|
| Full SaaS | Cloud | Cloud | Cloud | Teams comfortable with cloud-based log processing |
| Hybrid | On-premise | On-premise | Cloud | Sensitive environments, regulated industries, data sovereignty requirements |
Both models deliver identical platform capabilities. The hybrid model adds the on-premise component without reducing functionality in the SaaS layer.
Full deployment documentation is available to Arrakis customers.