Operations | Monitoring | ITSM | DevOps | Cloud

Chris Watts is Head of Enterprise Engineering at OpenRouter, building infrastructure for AI applications. Previously at Amazon and a startup founder. As large language models become core infrastructure for more and more applications, teams are discovering a familiar challenge in a new context: you can't improve what you can't see.

SAN JOSE, Calif., March 23, 2026 — Tigera, the creator and maintainer of Project Calico, today announced a major expansion of its Unified Network Security Platform for Kubernetes, aimed at helping enterprises consolidate infrastructure and accelerate the migration of legacy workloads to cloud-native platforms.

Most teams know they are spending money on AI in production. Far fewer realize how much they are spending outside production. It’s easy to get lost as you evaluate which model has the best responses, is fast enough, and cheap enough to run in production. That is because the AI bill usually shows up as a giant blob. It is easy to see the total.

Engineering teams are shipping faster than ever. AI coding tools like Claude Code and OpenAI’s Codex have quietly removed some of the biggest friction points in the development cycle — and the result is that FinOps teams are being asked to keep up with a pace most practitioners haven’t fully reckoned with yet. That acceleration has a cost consequence. More shipping means more services, more experiments, more infrastructure spun up without review cycles.

Building AI services with large language models and agentic frameworks often means running complex microservices on Kubernetes. Observability is vital, but instrumenting every pod in a distributed system can quickly become a maintenance nightmare. OpenLIT Operator solves this problem by automatically injecting OpenTelemetry instrumentation into your AI workloads—no code changes or image rebuilds required.

Large language models don’t work in a vacuum. They often rely on Model Context Protocol (MCP) servers to fetch additional context from external tools or data sources. MCP provides a standard way for AI agents to talk to tool servers, but this extra layer introduces complexity. Without visibility, an MCP server becomes a black box: you send a request and hope a tool answers. When something breaks, it’s hard to tell if the agent, the server or the downstream API failed.

In another post in this series, we discussed how to instrument large language model (LLM) calls. This can be a good starting point, but generative AI workloads increasingly rely on agents, which are systems that plan, call tools, reason, and act autonomously. And their non‑deterministic behavior makes incidents harder to diagnose, in part, because the same prompt can trigger different tool sequences and costs.