91 field notes on running agents
Notes from inside the agent loop.
Engineering reports on traces, evaluations, production failures and the human decisions around them.
journal · latest runs streaming
Latest
Fresh off the loop.
Start with a question
Pick the problem you have today.
01 / 04 · 61 articles
engineering
02 / 04 · 17 articles
platform comparison
03 / 04 · 8 articles
observability
Topic
engineering
The full trace
Every article. Longer bar, longer read.
- How to Trace LangGraph Agents: Nodes, Tools and Model Calls13 min
- How to Track LLM Token Usage Per Request, Run and Release14 min
- How to Estimate AI Agent Costs Before You Launch20 min
- AI Agent Evaluation in Production: A Practical Debugging Framework14 min
- The AI infrastructure stack: What each layer actually does17 min
- Data Quality in Machine Learning: A Practical Framework for Production AI17 min
- Enterprise AI adoption trends: A governance framework that actually ships15 min
- The AI Monitoring Toolchain: Wiring Tools Into One System17 min
- AI Observability vs MLOps: Where Each Discipline Starts and Stops22 min
- Cloud-Native AI Tools: Choosing What to Manage and What to Own15 min
- Instrumenting AI agents for observability16 min
- Monitoring vs observability: What actually explains an AI agent failure15 min
- AI Agent Audit Trails: Building Compliance-Ready Evidence for SOC 2 and GDPR17 min
- Voice AI Agent Observability: Tracing Latency, Turn-Taking, and Failures in Real-Time Voice Agents21 min
- LLM Gateway and Model Routing: A Decision Framework for Multi-Provider AI Systems21 min
- Semantic Caching for LLM Applications: When It Saves Money and When It Breaks Quality17 min
- AI Agent Cost Attribution: Chargeback and Showback for Multi-Tenant AI Spend20 min
- AI Agent Drift Detection: How to Detect Behavioral Changes Before Users Report Them18 min
- AI Agent Evaluation Datasets: How to Build Test Cases From Real User Interactions21 min
- AI Agent Guardrails: A Layered Safety Model for Tool-Using Systems23 min
- LLM-as-a-Judge Evaluation: When It Works, When It Fails, and What to Measure17 min
- AI Agent Release Checklist: What to Verify Before Changing a Prompt, Model, or Tool18 min
- AI Agent Cost Per Task: How to Measure Unit Economics Beyond Token Usage19 min
- AI Agent Failure Taxonomy: 12 Failure Modes Every Production Team Should Track27 min
- AI Agent Observability vs Traditional APM: What Changes and What Stays the Same?20 min
- AI Agent SLOs and Error Budgets: A Practical Reliability Framework21 min
- AI Agent Trace Schema: What to Capture for Debugging, Evaluation, and Cost Analysis19 min
- AI Observability Metrics That Matter: Latency, Cost, Quality, and Reliability12 min
- AI Observability Best Practices: A Practical Rollout Guide12 min
- AI Observability Dashboard: What to Track and How to Design It12 min
- Generative AI Observability: Monitoring Output Quality at Scale12 min
- Agentic AI Observability: Tracing Reasoning, Tools, and Autonomy Failures13 min
- AI Observability Platform: What to Require Before You Buy13 min
- AI Observability Tools: How to Evaluate and Choose13 min
- What Is AI Observability? A Practical Definition for Production Teams13 min
- AgentOps vs Tracify: Auto-Instrumentation vs Integrated Evaluation27 min
- Arize vs Tracify: LLM Monitoring vs Agentic Observability24 min
- Braintrust vs Tracify: Evaluation-First vs Integrated Eval25 min
- Datadog vs Tracify: Enterprise APM vs Specialist Agent Observability22 min
- Laminar vs Tracify: Trace Compression vs Comprehensive Coverage22 min
- Langfuse vs Tracify: Open-Source Tracing vs Comprehensive Evaluation24 min
- LangSmith vs Tracify: The TCO Trap and All-In-One Value20 min
- LangWatch vs Tracify: Open-Source Cost Tracking vs Commercial All-In-One29 min
- Latitude vs Tracify: Agent-Specific Tracing vs Integrated Evaluation24 min
- New Relic vs Tracify: Full-Stack APM vs Agent-Native Observability20 min
- OpenObserve vs Tracify: Unified Infrastructure Observability vs LLM-Specific Observability19 min
- Parea AI vs Tracify: LLM Development vs Agent Observability28 min
- Portkey vs Tracify: LLM Gateway vs Native Observability25 min
- Vellum vs Tracify: LLM Engineering Platform vs Agent Observability Platform22 min
- WhyLabs vs Tracify: Drift Detection vs Integrated Agent Observability19 min
- Agent Deployment and Rollback Strategies27 min
- Multi-Agent Orchestration Patterns: A Practical Guide26 min
- Tool-Calling Reliability: Handling Failures, Retries, and Malformed Calls18 min
- AI Evaluation Metrics: What to Measure in Production26 min
- AI Workflow Automation: A Safe Production Pattern21 min
- AI Agent Frameworks: How to Choose for Production27 min
- LLM APIs in Production: Reliability, Cost, and Observability24 min
- Vector Databases for AI Applications: A Practical Guide22 min
- Prompt Engineering for Production AI Systems22 min
- How to Build a RAG Chatbot That You Can Debug22 min
- Agent-to-Agent (A2A) Protocols Explained22 min
- AI Agent Security: A Practical Production Checklist22 min
- Context Window Management for Long-Running Agents18 min
- When the Agent Breaks at 3am: An Incident Response Guide for AI Systems22 min
- AI Agent Monitoring: Alerts That Lead to a Decision20 min
- How to Write a Tracify Blog Post Like a Pro22 min
- Prompt Versioning in Production: Connect Changes to Outcomes23 min
- Browser Agent Observability: Trace the User Journey, Not Just the Clicks23 min
- AI Agent Production Checklist: Before You Turn It On23 min
- AI Agent Regression Testing: Turn Production Failures Into Coverage20 min
- Human-in-the-Loop AI Agents: Design the Approval Boundary23 min
- OpenTelemetry for LLM Tracing: A Practical Starting Point22 min
- MCP Observability: Trace Tool Calls Across Agent Servers23 min
- AI Agent Memory: What to Store and What to Forget23 min
- AI Agent Cost Optimization: A Trace-Based Method23 min
- LLM Hallucination Detection: A Practical Production Playbook22 min
- LLM Latency Optimization: Find the Slowest Step First20 min
- Structured Outputs for AI Agents: A Safer Tool-Calling Pattern23 min
- Prompt Injection Defense for Tool-Using AI Agents24 min
- RAG Evaluation: How to Measure Retrieval and Answer Quality20 min
- AI Agent Architecture: The Production Building Blocks23 min
- Building Production-Ready AI Agents20 min
- AI Agent Testing: From Unit Tests to Production Evals20 min
- Prompt Versioning and Prompt Management27 min
- AI Agent Reliability: Failures, Retries, and Guardrails20 min
- LLM Tracing Explained20 min
- How to Reduce LLM Costs Without Hurting Quality27 min
- AI Agent Evaluation: A Practical Guide22 min
- How to Debug AI Agents in Production20 min
- LLM Observability: Metrics That Actually Matter23 min
- AI Agent Observability: The Complete Guide15 min