Schema on split
data architecture
A pipeline rule: land raw, split deterministically, enrich point-in-time, then shape for consumers.
kotona.app
Juha Huotari
Solution analyst in Helsinki. I work on data platforms, integration-heavy systems, and infrastructure that benefits from being slightly harder to misuse.
This is a field notebook for systems that have to keep working: data platforms, infrastructure, contracts, and agent-assisted engineering. Projects explain what exists. Notes keep the decisions, failures, and evidence worth reusing.
Start here
Four ways into the work: data, operations, agent workflows, and applied exploration.
Moving a live cluster to a new subnet
cluster operations
Migrating a Talos cluster from a legacy subnet to its own VLAN while it kept serving. The design was mostly an ordering problem, plus one bootstrap loop: the network controller lives inside the network it manages.
Judge agents by the next prompt
agent workflow
New agents carry more work from intent to evidence, but that changes rather than removes the supervision problem. The next prompt reveals whether the result needs repair or is ready to extend.
Delivered energy
energy systems
A business plan for battery-swap delivered energy that works in one narrow niche and matters most for what it implies about EVs, home batteries, and off-grid design.
Projects
All projectsBuilding an AgentOps substrate because agents kept editing the wrong line
Published and in active use
sprintctl and kctl began as schema-enforced sprint and knowledge tools. They now sit beside actionq, auditctl, and a live AgentOps cockpit, with each tool owning one kind of operational state.
Running a homelab cluster as a Git repository, not as muscle memory
Active operations repo
Appservice is the operations repository for a Talos-based Kubernetes cluster. Reconciliation, encrypted secrets, recovery gates, upgrades, and incident evidence live with the desired state instead of in operator memory.
Turning household analytics into an operating platform
Active household operating platform
Homelab Analytics keeps household reporting, planning, simulation, policy, trust, and agent-facing retrieval in one semantic model. Home Assistant is the device-facing partner, not the system of record for household reasoning.
Recent notes
All notesThe ref nobody adds
agent workflow
The sprint database held 831 work items and seven doc refs, six of them the same document. The reference command existed the whole time. The missing piece was never a tool — it was placement on the path the agent actually walks.
Layering you can't violate by accident
software architecture
homelab-analytics enforces its layer boundaries with pytest. A documented architecture is a request; a tested one is a constraint.
NFC tokens pointing at a manifest
hardware contracts
Cassettes, diskettes, and cartridges as physical handles for digital media. The rule that makes the whole family of projects work: the tag points, it never hosts.
Subprocess, not service
agent workflow
actionq-dispatcher runs agent work as subprocesses inside a per-invocation coordinator. The rejected alternative was a long-running agent service, and the rejection is most of the design.
The aftertaste of resolution
agent workflow
Conversational refinement produces the feeling of having thought clearly, which is not the same thing as having thought clearly. The fix was not discipline. It was writing the stop rule into the counterpart.