Practical articles on network and infrastructure automation — from CMDB design with Nautobot, to Ansible and Nornir workflows, to integrating AI/LLMs into the day-to-day life of a network engineer.
uv + ruff toolchain.uv replacing venv, fixing WSL crashes, VMware Fusion shared folders, dev-container patterns.Browse the archive below or use the tag index to filter by topic.
Found a typo, want to suggest a topic, or have a real-world automation problem you'd like to see covered? Open an issue or PR on GitHub, or reach out on LinkedIn or Bluesky.
AIOps pipelines need more than a generic Python image. They often need LLM clients, prompt tests, structured output validation, RAG evaluation tools, telemetry libraries, and repeatable replay datasets. The bsmeding/aiops_cicd_* images provide a ready-to-use CI base for those workflows.
If you want a repeatable way to deploy LibreChat, an Ansible role is easier to maintain than hand-editing Docker Compose files on every host. The bsmeding.librechat_docker role wraps the upstream stack and gives you consistent variable-driven configuration.
Python network automation often starts as a script and slowly becomes a change platform. Before that happens, it needs tests, dry runs, and guardrails.
Source-of-truth automation is powerful because it changes the system that other automation trusts. That also makes it risky. A broken sync job can create bad devices, duplicate prefixes, wrong interface names, or stale custom fields.