Metrickit

This gives Claude the knowledge to help you implement MetricKit for collecting real-world performance data and crash diagnostics from production iOS and macOS apps. You'll want this when you're setting up MXMetricManager subscribers, processing daily metric payloads for CPU and memory stats, handling diagnostic payloads with crash call stacks, or building telemetry pipelines. The documentation covers the complete flow from registration through JSON export, including the tricky bits like extended launch measurement and custom signpost metrics. One thing to know: MetricKit delivers data once per callback, so you need to persist the JSON immediately or risk losing it.