GTP SAP Aegis

Platform Overview

Four Integrated Components, One Unified Platform

Aegis is built on Python (FastAPI), Prometheus, Grafana, Docker / Kubernetes, and an integrated AI engine — covering every layer of SAP observability and protection.

Aegis Core

Metric ingestion layer

Python FastAPI · Prometheus exposition format. Collects CPU, memory, disk, SAP work processes, background jobs, and HANA metrics.

Aegis Insight

AI engine — anomaly detection & predictive analytics

Python · scikit-learn · Isolation Forest. Continuous anomaly scoring, 15/60/240-minute forecasts, cross-metric correlation.

Aegis Shield

Alerting & automated remediation

Prometheus Alertmanager · alert rules. Email, Slack, and PagerDuty routing with tiered severity — critical, warning, pre-emptive AI alerts.

Aegis View

Dashboard & visualisation layer

Grafana · pre-built SAP dashboards. Production monitoring dashboard plus AI engine anomaly, forecast, and baseline panels.

AI Engine

Aegis Insight — Predictive Intelligence for SAP

Aegis Insight connects to the Prometheus time-series store and continuously analyses ingested SAP metrics, surfacing anomalies, predicted threshold breaches, and root-cause signals before they become incidents.

Isolation Forest anomaly detection — flags statistical outliers across CPU, memory, HANA, and job metrics

Predictive alerts — linear and ARIMA forecasting for 15, 60, and 240-minute horizons with pre-emptive Aegis Shield alerts

Cross-metric correlation — links co-occurring anomalies (e.g. CPU spike + HANA connection surge) into a single incident context

Baseline learning — 7-day rolling-window baseline adapts to daily and weekly SAP load patterns

Insight API — FastAPI sub-application exposing prediction endpoints, anomaly scores, and correlation graphs as JSON

Aegis Insight — AI Capabilities

Anomaly Detection

Isolation Forest · contamination threshold configurable (default 0.15)

Predictive Forecasting

ARIMA + linear regression · 15 / 60 / 240 min horizons with confidence bands

Root-Cause Correlation

Pearson cross-metric correlation · incident grouping table

Baseline Learning

7-day rolling window · ±2σ bands overlaid on actual metric values

SAP Coverage

Comprehensive SAP & HANA Metric Collection

Aegis Core ingests metrics across every critical SAP and HANA layer — from OS-level CPU and memory through to SAP work processes, background jobs, and HANA service internals.

CPU: total usage, per-core, and load averages (1m / 5m / 15m)

Memory: RAM utilisation, SAP extended memory, roll memory, swap

Disk: filesystem usage, I/O throughput, latency, and IOPS

SAP Jobs: failed, cancelled, long-running, active, and scheduled (SM37)

Work Processes: total, busy, and free by type (dialog, background, update)

HANA: service memory, host CPU, data/log volume usage, connection count

Metric Categories

CPU & Load6 metrics

Memory & Swap6 metrics

Disk I/O5 metrics

SAP Background Jobs5 metrics

SAP Work Processes3 metrics

HANA Database5 metrics

Instance Reachability1 metric

Alerting & Remediation

Aegis Shield — Intelligent Alerting & Routing

Aegis Shield combines Prometheus Alertmanager with AI-driven pre-emptive alerts from Aegis Insight, delivering tiered notifications to Email, Slack, and PagerDuty based on severity — so critical issues never get lost.

11 pre-configured SAP alert rules covering CPU, memory, disk, jobs, work processes, and HANA

Tiered severity routing — critical alerts escalated to PagerDuty, warnings to Slack

Pre-emptive Aegis Insight forecast alerts — raised before a metric breaches its threshold

App & web service probing — HTTP/HTTPS endpoint and TCP port availability via Blackbox Exporter

TLS certificate expiry alerts — 14-day warning and 3-day critical thresholds

Alert Rules Reference

SAP Instance DownCritical

CPU > 95% for 2 minCritical

CPU > 85% for 5 minWarning

Memory > 95% for 2 minCritical

Disk > 90% for 2 minCritical

SAP Jobs FailedWarning

Long-Running Jobs > 60 minWarning

Dialog Work Processes < 2 freeCritical

HANA Memory > 85% for 5 minWarning

Visualisation Layer

Aegis View — Pre-Built SAP Grafana Dashboards

Two auto-provisioned Grafana dashboards give your team instant, deep visibility into SAP and HANA health — including AI-powered anomaly scores, 60-minute predictive forecasts, and cross-metric correlation graphs.

SAP Monitoring — Production: Instance health KPIs, CPU, memory, disk, work processes, background jobs, HANA metrics

Aegis Insight AI Dashboard: anomaly score gauges, 60-minute forecast panels with confidence bands, correlation group table

7-day rolling baseline overlays with ±2σ bands for CPU and memory

Branded with "Powered by Global Tech Professionals" pinned at the top of every dashboard

SAP Monitoring Dashboard

UP

Instance Health

0.07

Anomaly Score

42%

CPU Utilisation

0

Failed Jobs

78

Memory

65

Disk

18

Dialog WP

12

HANA Conn.

Deployment

From Demo to Production in Minutes

Aegis ships as a fully containerised stack (Docker Compose and Kubernetes). A zero-credential demo stack with simulated SAP metrics lets you evaluate the full platform without a live SAP system.

Zero-config demo stack — simulated SAP/HANA metrics included, no credentials required

Production Kubernetes deployment — one exporter per SID, multi-SID supported via Prometheus labels

Built-in Admin Panel at /admin/ — CRUD host inventory, subsystem configuration, live HANA connection testing

Role-based access control — superadmin and read-only guest roles with PBKDF2-hashed passwords and signed session cookies

Security hardening — non-root container (UID 1001), read-only filesystem, Kubernetes NetworkPolicies

Quick Start — Demo Stack


                  # Start the full demo stack

                  docker compose -f docker-compose.demo.yml up --build

Services Available

:3000Aegis View — Grafana Dashboards

:8000Aegis Core — Metrics Exporter & Admin

:9090Prometheus

:9093Alertmanager

Key Capabilities

What Makes GTP SAP Aegis Unique

AI-Powered Anomaly Detection

Isolation Forest model continuously scores every SAP metric and surfaces outliers before they escalate to incidents.

Predictive Forecasting

ARIMA and linear regression forecasts at 15, 60, and 240-minute horizons — raise alerts before thresholds are breached.

Zero-Config Demo Mode

Evaluate the entire platform in under a minute with simulated SAP/HANA metrics — no SAP credentials needed.

Enterprise Security

Non-root containers, read-only filesystem, PBKDF2 password hashing, HMAC-signed sessions, and Kubernetes NetworkPolicies.

Docker & Kubernetes Ready

Full Kubernetes manifests included — deploy one exporter per SAP SID and scale horizontally across landscapes.

Pre-Built Grafana Dashboards

Auto-provisioned SAP Monitoring and Aegis Insight dashboards — no manual dashboard setup required.

Built-In Admin Panel

Manage SAP/HANA host inventory, subsystem collectors, and live connection tests — all from the browser without editing config files.

Multi-Channel Alerting

Email, Slack, and PagerDuty receiver configs included — tiered by severity with group-by SID and alertname routing.

Platform Architecture

Aegis ingests metrics via RFC / OData / HANA SQL, stores them in Prometheus, and feeds three parallel downstream components — Aegis View, Aegis Insight, and Aegis Shield.

SAP System

ABAP / HANA
RFC · OData · HANA SQL

Aegis Core — Metric Ingestion Layer

Python SAP Metrics Exporter (FastAPI) · :8000/metrics · Prometheus exposition format

▼ Prometheus (scrape + storage) :9090 ▼