Surgical InnoDB Buffer Pool tuning, high-availability replication (InnoDB Cluster / Group Replication), live SQL profiling, and direct expertise from the MySQLTuner co-author.
Open Source Tools, Scripts & Platform Engineering Accelerators
Lightpath's active contribution to Open Source sovereignty: storage engine diagnostic scripts, declarative blueprints (GitOps, Kubernetes), and engineering accelerators to audit, secure, and scale your data platforms without vendor lock-in.
Production-Ready Engineering Accelerators
Field-proven utilities to optimize latency, enforce CIS benchmarks, and orchestrate databases on Kubernetes.
MySQLTuner-perl
The global standard Perl diagnostic script for live MySQL, MariaDB, and Percona Server instances (co-developed by Jean-Marie Renouard). Audits Buffer Pool allocation, lock queues, index usage, and generates instant tuning heuristics.
perl mysqltuner.pl
CIS Benchmarks Tools Suite
Automated security compliance audit engine (100% Python Standard Library—zero external dependencies, Air-Gapped ready) authored by Jean-Marie Renouard. Evaluates 18 official CIS Benchmarks & DISA STIG baselines (887 controls) across MySQL, PostgreSQL, MariaDB, MongoDB, Cassandra, and RHEL 8/9/10 via Local, SSH, or Docker.
pip install cis-benchmarks-tools && audit-cis --auto-detect
Linux Hardening & Security Guide v2.0 — Production Baseline & Defense-in-Depth
The authoritative enterprise Linux hardening manual published by LightPath and authored by Jean-Marie Renouard. Complete 2026 engineering overhaul to fortify mission-critical database hosts and production servers: RHEL 9 / CentOS Stream 9 / Rocky Linux 9 and Ubuntu 24.04 LTS baselines, unified nftables & fail2ban firewalling, hardware-backed FIDO2 SSH MFA (ed25519-sk), strict systemd unit sandboxing (ProtectSystem=strict, Seccomp-BPF), SELinux/AppArmor enforcement, rootless Podman containers with Trivy CVE scanning, LUKS2 encryption with Argon2id, and SHA-512 AIDE file integrity.
In-Database AI Frameworks & Production Blueprints
Architectural patterns, extensions, and automation scripts authored by Jean-Marie Renouard to execute data processing models (embeddings, classification, vector search) directly inside mission-critical database engines (PostgreSQL, MariaDB). Eliminates sensitive data egress to external cloud APIs, cuts inference infrastructure costs by over 80%, and guarantees sub-millisecond latencies with full DORA and GDPR auditability.
CREATE EXTENSION vector; SELECT title, vector_cosine(embedding, query_embedding) FROM articles ORDER BY 2 DESC LIMIT 5;
Kubernetes DBaaS GitOps Blueprints
Declarative reference blueprints to orchestrate cloud-native database clusters (CloudNativePG, NoSQL operators, Valkey) on Kubernetes with automated backups, self-healing, and zero configuration drift.
Grafana Tools Configurator
Utilities and pre-built templates to rapidly generate production-grade Grafana dashboards, correlating Linux host telemetry with database query latencies.
MongoDB Ecosystem & Drivers
A technical reference guide on drivers, administration tooling, collection indexing, and scaling strategies for production MongoDB clusters under high volume.
Open Source Migration Framework (Zero Lock-in)
Evaluation framework and transition checklists to liberate operations from proprietary licenses (Oracle, SQL Server) and migrate smoothly to PostgreSQL, MariaDB, Valkey, and ScyllaDB.
Spotlight: CIS Benchmarks Tools Suite (18 Benchmarks • 887 Controls)
An enterprise-grade automated security compliance engine engineered with zero external dependencies (Python Standard Library only) to evaluate databases and Linux hosts against official CIS (Center for Internet Security) Benchmarks and DISA STIG guidelines.
Architecture & Engineering Principles
-
100% Python Standard Library (PSL ONLY — Air-Gapped) Zero third-party pip dependencies (no requests, jinja2, paramiko, or pyyaml). Guaranteed standalone execution on any standard Python 3.8+ runtime in banking, sovereign, or air-gapped environments.
-
4-Tier Hybrid Execution Context Engine Agentless auditing across 4 topologies: (1) Local Machine (--local), (2) Remote SSH (--ssh user@host), (3) Local Docker Container (--docker), and (4) Remote Docker Container over SSH (--ssh user@host --docker), plus service auto-detection (--auto-detect).
-
5-State Compliance Taxonomy & Telemetry Integrity Strict 5-state classification (Pass, Fail, technical Error isolated from security failures, Manual with operational runbooks, Not Applicable), full stderr preservation without 2>/dev/null masking, decoupled JSON rules (rules/*.json), and automatic secret masking (***MASKED***).
-
100% Offline SVG Reports & SIEM / DevSecOps Exports Native bilingual HTML reports (--lang fr|en) with pure inline SVG Donut charts, JSON exports for SIEM ingestion (Splunk, ELK), XML (JUnit/STIG for GitLab CI, GitHub Actions, Jenkins pipelines), and TXT.
pip install cis-benchmarks-tools && audit-cis --auto-detect
python3 audit_cis.py --target postgresql17 --ssh admin@db-prod --format html
python3 audit_cis.py --target mysql-community84 --docker mysql84 --format json
18 Supported CIS & DISA STIG Benchmarks & Live Demo Reports
Each benchmark is powered by a decoupled JSON rule specification (rules/*.json) and generates a self-contained HTML report with dark mode and inline SVG charts.
PostgreSQL 16 / 17 / 18
postgresql16— CIS Benchmark v1.1.0 (71 controls)postgresql17— CIS Benchmark v1.1.0 (71 controls)postgresql18— CIS Benchmark v1.0.0 (71 controls)
MySQL 8.0 / 8.4 LTS / 9.7
mysql80— CIS v1.5.0 (70 controls)mysql-community84— CIS v1.1.0 (79 controls)mysql-enterprise84— CIS v1.1.0 (70 controls)mysql-community97 / enterprise97— CIS v1.0.0 (2×70)
MariaDB 10.6 & 10.11 LTS
mariadb106— CIS Benchmark v1.3.0 (74 controls)mariadb1011— CIS Benchmark v1.0.0 (75 controls)- Galera Cluster hardening, PAM/TLS auth, audit plugin & filesystem permissions.
MongoDB 7.0 & 8.0
mongodb7— CIS Benchmark v1.2.0 (23 controls)mongodb8— CIS Benchmark v1.0.0 (23 controls)- SCRAM-SHA-256, x.509/TLS, ReplicaSet keyFile, RBAC roles & audit logging.
Apache Cassandra 4.0 / 4.1 / 5.0
cassandra40— CIS Benchmark v1.3.0 (20 controls)cassandra41— CIS Benchmark v1.0.0 (20 controls)cassandra50— CIS Benchmark v1.1.0 (20 controls)
RHEL 8 / 9 / 10 (CIS & STIG)
rhel8— CIS v4.0.0 & DISA STIG v2.0.0 (20 controls)rhel9— CIS v2.0.0 & DISA STIG v1.0.0 (20 controls)rhel10— CIS Benchmark v1.0.1 (20 controls)
Spotlight: In-Database AI — Rational Processing & Data Sovereignty
Execute AI models, vector search, and classification pipelines directly within the database storage engine—without sensitive data egress to external third parties.
4 Architectural Pillars & Enterprise Guarantees
-
Native Vector Search (pgvector & pg_embedding) High-performance HNSW and IVFFlat indexation, sub-millisecond cosine similarity co-located with transactional records without external ETL pipelines.
-
Zero Data Egress & Total Sovereignty (DORA & GDPR) Process mission-critical sensitive records (PII, banking transactions, healthcare data) in situ. Zero reliance on external opaque cloud LLM endpoints.
-
FinOps Optimization & Zero Unnecessary GPU Costs Save up to 80% on AI infrastructure expenditure by leveraging existing database compute nodes without dedicated multi-GPU clusters.
-
Model Context Protocol (MCP) Connectors & Traceability Standardized MCP gateways featuring granular access control, dynamic data masking, and cryptographic auditability for all analytical queries.
4 Production Deployment Blueprints
PostgreSQL & pgvector RAG
High-velocity semantic search and Retrieval-Augmented Generation (RAG) hosted directly inside the relational engine with zero ETL overhead.
In-Database Classification
Automated transactional flow categorization and routing co-located with storage via stored procedures and embedded light models.
Audit & Anomaly Detection
Real-time query behavioral analysis and anomaly scoring to preempt production outages, data breaches, and malicious patterns.
Connecteurs MCP Sécurisés
Model Context Protocol (MCP) enterprise gateway architecture empowering AI agents to query database assets securely with strict audit trails.
Mastery of the Open Source & Distributed Database Ecosystem
Explore the 12 database engines we architect, secure, and tune in mission-critical production. Filter by family or swipe through the carousel.
Platform Engineering & DevOps Automation Stack
From Ansible, Terraform, and OpenTofu to ArgoCD, Jenkins, Rundeck, Ansible Tower, GitLab CI, and GitHub Actions: we industrialize the entire lifecycle of your data platforms.