SentinAI Docs

SentinAI Overview

guide/overview.md

← Home

SentinAI Overview

What is SentinAI?

SentinAI is an Autonomous Node Guardian — a Next.js dashboard for monitoring and auto-scaling Optimism-based L2 networks.

Key Features

Real-Time Monitoring

  • L2 Block Monitoring: Real-time block height display and network metrics
  • K8s Integration: AWS EKS connection with cached dynamic token generation (10-minute expiry) for low-latency polling

Intelligent Scaling

  • Dynamic Resource Scaling: Hybrid auto-scaling engine using CPU, TxPool, and AI insights
  • Predictive Scaling: AI-powered time-series analysis (Claude Haiku 4.5) predicts optimal vCPU/MEM allocation 5 minutes ahead
  • Adaptive Tiers: Idle (<30) → 1 vCPU | Normal (30-70) → 2 vCPU | High (>70) → 4 vCPU

AI-Powered Analysis

  • Log Anomaly Detection: Claude-based analysis for Optimism Rollup components (op-geth, op-node, op-batcher, op-proposer)
  • Model Benchmarking: Compare AI model performance (Qwen, Claude, GPT, Gemini) with latency, cost, and accuracy metrics
  • Tier-Based Model Selection: Fast Tier (qwen3-80b-next, 1.8s) and Best Tier (qwen3-235b, 11s)

Testing & Simulation

  • Stress Test Simulation: Simulate peak load scenarios (8 vCPU / 16 GiB)
  • Safety Mechanisms: Cooldown periods and simulation mode by default

Architecture Highlights

Hybrid Scoring Logic (0-100)

  • CPU & Gas (60%): Real-time load indicators
  • TxPool (20%): Pending transaction bottleneck detection
  • AI Severity (20%): Proactive scaling based on log anomaly risks

AI Log Analysis Engine

  1. Holistic Context Window: Aggregates logs from 4 core components to detect complex cross-component issues
  2. Senior Engineer Persona: Prompted to check for security, consensus, and liveness issues
  3. Actionable Intelligence: Structured JSON output with severity, summary, and action items grounded in official Optimism documentation

Safety First

  • Cooldown: 5-minute freeze after scaling to prevent flapping
  • Simulation Mode: Dry-run execution by default for safety
  • AWS Integration: Standard AWS credential chain (env vars, ~/.aws/credentials, or IAM Role)

Supported Networks

  • Optimism / OP Stack

For setup instructions, see Setup Guide.
For hands-on demos, see Demo Scenarios.