Total Traces
0
events evaluated
Hallucinations
0
warning + critical flags
Self-Heals
0
prompt mutations deployed
Escalations
0
low-confidence flags
Pass Rate
—
info-severity / total
Critical Rate
—
critical-severity / total
Evaluator Heatmap
| Evaluator | Runs | Fails | Avg Score | Crits |
|---|---|---|---|---|
| Loading… | ||||
Severity Timeline
No trace data yet
info
warning
critical
Query Type Breakdown
| Query Type | Count | Fail Rate |
|---|---|---|
| Loading… | ||
Agent Improvement — care-advisor-v2
Run scenarios in Synthetic Response mode, let a heal deploy, then re-run in Live Generation mode to see the before/after.
Live Traces
Waiting for SSE alerts
Self Healing
No pending candidates
Auto-approve is on — healed prompts deploy instantly and appear in History
No healing history yet
Test & Debug MCP Servers
Ready. Ask anything about your Phoenix observability data — spans, prompts, datasets, experiments. I'll query the live MCP server to answer.
Activity Log
Ready — waiting for ADK / MCP events