Observability at scale
Grafana, Loki, Thanos, Vector, Elasticsearch. Stack audits, ingestion tuning, Fluentd-to-Vector migrations, recording rules, query performance, and SLO dashboarding.
- Grafana
- Loki
- Thanos
- Vector
RetakeData / infrastructure practice
Take control of your data
RetakeData helps infrastructure teams regain control of sensitive systems — observability, automation, virtualization, private cloud, and local AI. Designed and operated by a senior SRE with 10+ years on high-stakes platforms.
Hardware to software · Full-stack infrastructure · Bare metal to private AI
FULL-STACK CONTROL
01
What we do
On-prem, private-cloud, and hybrid infrastructure engineering. Observability, automation, virtualization, and private AI — for teams that need to keep sensitive workloads under their own control.
Infrastructure automation & IaC
Terraform multi-provider modules, Ansible, GitOps delivery pipelines, Consul and NetBox inventory. The automation that lets small teams operate at fleet scale.
- Terraform
- Ansible
- NetBox
- Consul
Proxmox / Ceph / on-prem HA
100+ Proxmox nodes deployed with PXE automation. Storage backends including ZFS, NFS, SAN, NVMe-oF, and Ceph. Migrations from vSphere, HA cluster design.
- Proxmox
- Ceph
- ZFS
- HA
02
Private AI for infrastructure teams
Some operational data should not leave your network: incidents, logs, runbooks, internal docs, and procedures. RetakeData builds local AI systems for those environments: vLLM model serving, private RAG pipelines, RBAC-aware assistants, and integrations with your existing Proxmox, observability, and documentation stack. No external API dependency for sensitive workflows, predictable costs, full data control.
OSS Models
vLLM
Llama
Qwen
Mistral
Embedder
BGE-M3
Nomic
sentence-transformers
Vector DB
pgvector
Qdrant
Chroma
03
Selected work
A few receipts from 10 years operating infrastructure that can't quietly fail.
Observability at high-stakes scale
Operated multi-cluster observability handling 11 TB/day across Loki, Elasticsearch, and Thanos. Fine-tuned ingestion, migrated Fluentd to Vector, built recording rules converting TB/day of load-balancer logs into metrics. In a business where minutes of downtime mean seven-figure impact.
VM delivery pipeline
Built a 6-phase autonomous VM delivery pipeline: Git PR, Terraform across vSphere/Proxmox/OpenStack, Consul and NetBox auto-registration, Ansible configuration, HAProxy backend registration, Centreon monitoring. No Kubernetes needed.
On-prem AI infrastructure
Deployed GPU-equipped servers in the datacenter running vLLM with a private RAG pipeline over 1000+ documents. Built Graphia, an RBAC-aware SRE agent that abstracts Grafana complexity for engineering teams.
04
Open Source
Tools built around real operational pain points, kept practical, open, and useful beyond our own environment.
OPEN SOURCE
SSHplex
Modern SSH multiplexing with multi-source inventory and tmux or iTerm2 backends.
GitHub →
OPEN SOURCE
OpenClaw Audit TUI
Terminal audit UI for OpenClaw sessions with live events and real-time streaming.
GitHub →
OPEN SOURCE
terraform-provider-centreon
Terraform provider for Centreon API V2, monitoring configuration managed as infrastructure as code.
GitHub →Need to keep critical infrastructure under your control?
If your team is building on-prem, private-cloud, or hybrid infrastructure for sensitive workloads, let's talk.
Get in touch