Alexandre Chappaz

Platform and observability engineer at Consensys, working on Web3 infrastructure — the Kubernetes platform our products run on, and the observability stack that keeps it honest.

Tech I work with

• Orchestration & GitOps — Kubernetes, ArgoCD (server-side diff, app-of-apps), Crossplane control planes, External Secrets Operator, Helm, Kustomize, Jsonnet
• Observability — Grafana, Loki, Mimir, Tempo, Pyroscope, Alloy, OpenTelemetry, Prometheus, SLOs and multi-window burn-rate alerting
• Cloud & infra — AWS (EKS, IRSA), Cloudflare, Terraform, GitHub Actions
• Code — Go, TypeScript, Python

At Consensys

• ask-o11y — AI observability assistant for Grafana
• Web3 Factory — internal multi-tenant Kubernetes platform

Open source

• mcp-grafana — MCP server for Grafana
• grafana-llm-app — LLM extensions for Grafana
• plugin-tools — tooling for building Grafana plugins

Personal projects

• rkllama — Ollama alternative for Rockchip NPUs: run LLMs locally with optimized NPU support on RK3588-class devices
• github-actions-cache-server — self-hosted cache server, compatible with the official actions/cache
• plugin-validator — validation tool for Grafana community plugins
• homelab — GitOps-managed Kubernetes cluster (ArgoCD + Jsonnet), from ingress to media automation

Currently into

• LLM inference on edge hardware — Rockchip NPUs, quantized models, llama.cpp forks

---