Day-2 Operational Model for CalypsoAI Inference Defend
Overview
This document outlines the Day-2 Operational Model for managing and sustaining the CalypsoAI Inference Defend solution in an on-premises deployment hosted within a customer’s cloud environment using Kubernetes. It is designed to help platform owners, DevOps teams, and security stakeholders understand the ongoing responsibilities required to maintain system health, manage scanner performance, and ensure secure operations post-deployment. The model includes key sustainment activities, recommended ownership, and operational frequencies, providing a structured framework to define internal support and sustainment processes. This guide also highlights optional CalypsoAI support services available to assist with scanner tuning, updates, and operational optimization.
Deployment Context
- Environment: Customer-hosted AWS infrastructure
- Orchestration: Kubernetes (EKS or self-managed)
- Solution: CalypsoAI Inference Defend (security layer at inference level for GenAI systems)
Key Areas of Sustainment
|
Category |
Description |
Owner(s) |
Frequency |
|
Application Monitoring |
Monitor pod health, restarts, latency, and logs via CloudWatch, Prometheus, etc. |
DevOps / SRE |
Daily / Continuous |
|
Scanner Performance |
Review scanner results, false positives, blocked prompt rates, and feedback loop configuration. |
Security / AI SME |
Weekly / As needed |
|
Policy & Scanner Tuning |
Update scanning policies or create custom scanners as use cases evolve. |
Security / Admin |
Monthly / Event-driven |
|
Model/Image Updates |
Pull and deploy updated scanner logic, model weights, or application images provided by CalypsoAI. |
DevOps / CalypsoAI |
Monthly / As released |
|
Integration Health |
Ensure APIs and integrations (e.g., to LLM backends or front-end interfaces) remain stable after changes. |
App Owner / DevOps |
Weekly / As needed |
|
License Usage Monitoring |
Track scan volume, feature usage, and ensure license compliance. |
Admin / Platform Owner |
Monthly |
|
Infrastructure Maintenance |
Maintain EKS or self-managed K8s cluster, apply security patches, manage DNS, load balancers, storage. |
DevOps / Cloud Team |
Weekly / As needed |
|
Access & Audit Controls |
Review user access controls, logs, and audit trails. |
Security / Compliance |
Quarterly |
|
Backup & Disaster Recovery |
Ensure configuration backups and restoration procedures are in place. |
DevOps / Platform Owner |
Monthly / Quarterly |
|
Support & Issue Escalation |
Triage issues, create support tickets with CalypsoAI as needed. |
Internal Support / Admin |
As needed |
Recommended Roles for Sustainment
- Platform Owner: Oversees deployment and integrations
- Security Analyst / AI Risk SME: Owns scanner tuning and risk policy
- DevOps / SRE: Maintains infrastructure and monitors health
- Support Liaison: Coordinates with CalypsoAI support when required
Optional CalypsoAI Support Package Inclusions
If opted in, our team can assist with:
- Quarterly health reviews
- Scanner update guidance
- Performance optimization support
- Troubleshooting assistance and fast-path escalation