Guides
Guides
Short guides for when an agent misbehaves, and for the record you keep of what it did. Each one answers the question first, then the steps.
Debugging
- What are the best practices for AI agent observability?
Give every run an id, log each step in order, and watch error rate, volume, and silence.
- How do you create an audit trail for AI agent actions?
Send one event per action. The trail is not tamper-evident, and retention tops out at 90 days.
- How do I debug my AI agent when it does weird things?
Record one run in order, then stop at the first step that is wrong.
- How do I catch an AI agent stuck in a loop?
Look for the same tool name on one run id. There is no loop detector.
Also on the site
- How to monitor AI agents
Record each run, then alert on error rate, volume, and silence.
- What is LLM observability?
The record of model calls, and how it differs from agent tracing.
- AI agent observability tools
What the tools record, and when another product is a better fit than AgentTracer.
- AI agent monitoring
Timeline, weekly digest, and anomaly alerts.
- Pricing
Free, Pro, and Team. Free does not ask for a card.