Under-the-hood metrics from Claude Code sessions — cost per session, cache hit rates, token usage, tool call counts, and per-phase success rates. Use this to monitor agent efficiency, spot regressions in API costs or cache performance, and identify which pipeline phases fail most often or cost the most.

Total Sessions i

--

Avg Cost i

--

Avg Cache Hit i

--

Avg TTFT i

--

Avg Tool Calls i

--

Avg Subagents i

--

Avg Duration i

--

Avg Turns i

--

Phase Effectiveness

Success Rate by Phase i

Phase Efficiency Summary i

Phase Sessions Success Rate Avg Cost Avg Duration Total Cost
Loading...

Cost by Phase

Token Usage by Phase

Cache Hit Rate Over Time

Duration by Phase

Success / Failure Rate

Top Tool Usage

API Request Latency Distribution

Tool Success Rate (OTel)

Sessions

Build Issue Phase Model Cost Duration Turns Cache % Tools Subagents TTFT Result
Loading...