03 / Stack & Analytics
Tools with a purpose.
A small, deliberate toolkit. Chosen for clear interfaces, repeatable delivery, and observable behavior.
Content inventory
- Reference architectures
- 02.
- Documents
- 03.
- Technologies
- 09.
Calculated from the local content at build time. These are document counts, not live infrastructure metrics.
Technology coverage
Number of reference architectures that use each technology.
- 2 / 2
2 / 2 - 1 / 2
1 / 2 - 1 / 2
1 / 2 - 1 / 2
1 / 2 - 1 / 2
1 / 2 - 1 / 2
1 / 2 - 1 / 2
1 / 2 - 1 / 2
1 / 2 - 1 / 2
1 / 2
The operating model
Infrastructure / declare the desired state
Terraform describes the foundation. Kubernetes provides a consistent runtime, while Argo CD reconciles application state. Keep ownership boundaries explicit so two controllers never compete for the same resource.
Data / make correctness inspectable
Python handles ingestion, PostgreSQL defines transactional boundaries, and dbt makes transformations reviewable. Prefer replayable batches and explicit data contracts over invisible one-off fixes.
Observability / close the feedback loop
Prometheus captures operational signals; Grafana connects them to the questions an operator needs to answer. Every page-worthy alert should point to a decision and a runbook.