Agent Conversation sorting and filters by cost, tokens, and tool calls
Sentry has introduced new sorting and filtering capabilities within the Explore > Agents interface. Users can now organize conversation data based on specific metrics including total cost, token count, tool call frequency, and error occurrences.
Verified State Diff
Impact & Verification Analysis
Developers and engineering teams building and maintaining LLM-powered applications using Sentry's agent monitoring tools.
It allows for data-driven optimization of AI agents, enabling teams to identify high-cost or high-latency conversations and troubleshoot specific tool-use failures efficiently.
Full Fact Overview
This update enhances the observability layer for AI agents within the Sentry platform. By exposing granular telemetry data—specifically cost, token usage, and tool execution metrics—directly in the UI, Sentry enables developers to perform cost-attribution analysis and identify performance bottlenecks in LLM-integrated workflows. This shift moves the platform from simple error tracking toward comprehensive AI performance monitoring.