菜单
Compare and configure mobile APM vendors (Firebase Performance, Sentry Performance, Datadog RUM Mobile, Embrace, New Relic Mobile) and pick the right one for the app. Use when selecting or migrating a performance-monitoring stack.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Emit JSON-structured logs on mobile with correlation ids, consistent field names, and redaction. Use when setting up logging in Kotlin, Swift, Dart, or TypeScript mobile code.
Emit counters, gauges, and histograms from mobile apps using OpenTelemetry metrics, with patterns for batching, labels, and avoiding cardinality explosions. Use when adding product or engineering metrics beyond the core RUM set.
Design actionable metric alerts for mobile telemetry that deep-link to traces and sessions, avoid paging noise, and page only for symptoms users feel. Use when creating or cleaning up Datadog / Grafana / Sentry alerts.
Define and capture the core mobile RUM SLIs (startup time, jank, ANR, crash, network error rate, battery) with Firebase, Sentry, Datadog, or OTel. Use when setting up client-side performance metrics.
Measure mobile UX quality with frozen frames, slow frames, ANRs, app hangs, and interaction-to-next-paint style metrics. Use when diagnosing jank, adopting JankStats / MetricKit, or reporting UX health.
How mobile session replay works (view-hierarchy reconstruction, not screen recording), how to configure privacy masking, and the cost / privacy tradeoffs. Use when evaluating or enabling Sentry, Datadog, or LogRocket session replay on mobile.
| name | apm-mobile |
| description | Compare and configure mobile APM vendors (Firebase Performance, Sentry Performance, Datadog RUM Mobile, Embrace, New Relic Mobile) and pick the right one for the app. Use when selecting or migrating a performance-monitoring stack. |
"Mobile APM" overlaps with crash, RUM, and tracing. Picking one tool (or a small stack) depends on what you need to attribute -- user journeys, backend latency, or native crashes. Choose consciously; do not stack four overlapping SDKs.
FirebasePerformance.getInstance().isPerformanceCollectionEnabled = true
val trace = FirebasePerformance.getInstance().newTrace("home_feed_load")
trace.putMetric("items_returned", 42L)
See sentry-mobile.
Datadog.initialize(
appContext: .init(),
trackingConsent: .granted,
configuration: Datadog.Configuration.builderUsing(
rumApplicationID: "...", clientToken: "...", environment: "prod"
).trackURLSession(firstPartyHosts: ["api.example.com"]).build()
)
If you start greenfield, prefer OpenTelemetry for traces and metrics, then send to the vendor via OTLP. This keeps the wire format portable and lets you change vendors without re-instrumenting the app. See distributed-tracing-mobile.
tracesSampleRate = 0.15 and replaysSessionSampleRate <= 0.05.sessionSampleRate and sessionReplaySampleRate separately.Budget your monthly DAU * session/user * sample rate to estimate cost within 10% before rollout.
Whatever you choose:
symbolication).analytics-privacy).Moving between vendors:
init happens off the main thread before Application.onCreate completes: instability../gradlew :app:analyzeReleaseBundle, xcrun size Build/Products/Release-iphoneos/MyApp.app).