Agent Logs: Trace Tool Calls, Retries, and Outcomes
Connect agent runs to their tool calls, permissions, retries, parallel work, and final outcomes through a clear execution record.
THE SIGNAL / COLLECTION
AI & LLM.
Follow the details.
How model requests become observable application behavior. Explore event contracts, measured token usage, prompt privacy, and the relationship between agent tasks and tool attempts.
Where to begin
Start with the AI logging guide if you are defining an event contract. Continue with token measurement for usage questions, prompt logging for data-handling decisions, or agent tracing for workflows that take several steps.
Articles in AI & LLM
LLM Token Logging: Measure Usage, Latency, and Cost
Build an interpretable record of LLM usage, timing, retries, and cost estimates while preserving incomplete accounting states.
Prompt Logging Without Leaking Sensitive Data
Use immutable prompt versions, safe metadata, and tested redaction boundaries to investigate AI behavior without routine transcript capture.
AI Logs API: A Practical Guide to Useful Events
Design AI events that connect requests, model attempts, validation, and outcomes without filling logs with sensitive content.