Building Stateless AI Tool Gateways with FastMCP, the 2026 MCP Spec, and Kubernetes
How to build stateless AI tool gateways with FastMCP, the 2026 MCP specification, and Kubernetes, and why statelessness finally makes MCP scale.
The archive
669 posts on Apache Iceberg, lakehouse architecture, data engineering and applied AI.
How to build stateless AI tool gateways with FastMCP, the 2026 MCP specification, and Kubernetes, and why statelessness finally makes MCP scale.
Write-Audit-Publish with Apache Iceberg branches brings CI/CD to your data: stage, audit, then fast-forward to main so consumers never see unvalidated bytes.
A practical guide to modernizing a data warehouse to Apache Iceberg without downtime, using federation first, then redirecting new data, then.
Two open specifications, the Agentic Graph Specification and the Open Agent Profile, turn agent plans and agent identity into portable, reviewable files.
Arrow Flight and ADBC move database results as columnar data, ending the row-oriented bottleneck between engines and applications. Here's how.
Budgeting for agentic analytics when every question costs something different: token economics, query economics, instrumentation, and the cost controls.
The five layers of an agentic lakehouse and where the MCP server sits: storage, catalog, semantic layer, MCP gateway, and agent surface, plus identity.
Autonomous table optimization when query workloads stop being predictable: observing file layout and query patterns, scoring compaction work, adaptive.
How to build an Apache Iceberg lakehouse that runs fully offline: storage, catalog, compute, cross-zone transfer, compliance, and the failure modes that bite.
Wiring analytical queries to transactional APIs in closed-loop decision agents: conditional writes, sagas with compensations, decision records, and blast.