Add golden metrics and summary metrics for APM AI Agent

[msanthoshi-coder]

Summary

Adds golden_metrics.yml for the apm-ai_agent entity type with three key metrics:

• Throughput — invocations per minute, sourced from both OpenTelemetry spans (filtered by gen_ai.agent.name) and APM transaction duration events.
• Latency — average response time in milliseconds, from OTel span duration and APM transaction duration.
• Error rate — percentage of errored spans/transactions, using otel.status_code = 'ERROR' for OTel and APM error count ratio.
• Both OpenTelemetry and APM data sources are supported for each metric.

Adds summary_metrics.yml for the apm-ai_agent entity type, referencing the three golden metrics as summary metrics displayed on the entity overview:

• Response time (ms) — references responseTimeMs golden metric
• Throughput (requests per minute) — references throughput golden metric
• Error rate (percentage) — references errorRate golden metric

[entityBot]

⚠️ Review Required ⚠️

👋 This PR has seen no activity in over 3 days and is still awaiting approval from: @newrelic/AIM

Please review or provide an update.