What React Native libraries do you use?
React Navigation, Hermes, RN New Architecture, Expo (mobile only)
Are you using sentry.io or on-premise?
sentry.io (SaS)
Are you using any other error monitoring solution alongside Sentry?
Yes
Other Error Monitoring Solution Name
Datadog
@sentry/react-native SDK Version
8.10.0
How does your development environment look like?
react-native: 0.83.2
react: 19.2.0
Android: hermesEnabled true, newArchEnabled true
Android SDK API levels: 31, 35, 36
@react-native-firebase/messaging in use (FCM push)
Sentry.init()
Sentry.init({
dsn: '...',
tracesSampleRate: 0.2,
integrations: [
Sentry.reactNativeTracingIntegration(),
Sentry.reactNavigationIntegration(),
],
})Steps to Reproduce
1. Android app with app start tracking + FCM push.
2. Fully close the app.
3. Send a push → Android spawns the process in the background; the native SDK stamps the app start timestamp at process creation.
4. User opens the app some seconds later → first Activity created, first frame drawn.
5. Check measurements.app_start_cold for that session.
Expected Result
Cold start reflects the real user-perceived startup (process already resident, ~1s), or the measurement is dropped. Like the Cocoa SDK does for pre-warmed starts getsentry/sentry-cocoa#1896
Actual Result
app_start_cold is the full gap between background process creation and the first frame after the user opens the app.
Example production trace: Cold Start span = 17.2s, but its children are Process Initialization (181ms) and JS Bundle Execution (1037ms), with a ~16s gap of no activity in between (the process idling in background before the user opened it):
app.start.cold "Cold Start" [17233ms]
├─ "Process Initialization" [181ms] T+0.00s
│ ── ~16s idle, no spans ──
└─ "JS Bundle Execution" [1037ms] T+16.19s
Values range from ~4s to ~56s but never above 60s, matching the 1-minute cap in AppStartMetrics. This creates wrong values for p95/avg cold start on Android.
Looks like appLaunchedInForeground is only evaluated in onActivityCreated(), which fires when the user finally opens the app.
Related fixes I checked don't cover this:
The iOS equivalent was fixed by dropping pre-warmed starts: getsentry/sentry-cocoa#1896
What React Native libraries do you use?
React Navigation, Hermes, RN New Architecture, Expo (mobile only)
Are you using sentry.io or on-premise?
sentry.io (SaS)
Are you using any other error monitoring solution alongside Sentry?
Yes
Other Error Monitoring Solution Name
Datadog
@sentry/react-native SDK Version
8.10.0
How does your development environment look like?
Sentry.init()
Steps to Reproduce
Expected Result
Cold start reflects the real user-perceived startup (process already resident, ~1s), or the measurement is dropped. Like the Cocoa SDK does for pre-warmed starts getsentry/sentry-cocoa#1896
Actual Result
app_start_coldis the full gap between background process creation and the first frame after the user opens the app.Example production trace: Cold Start span = 17.2s, but its children are Process Initialization (181ms) and JS Bundle Execution (1037ms), with a ~16s gap of no activity in between (the process idling in background before the user opened it):
app.start.cold "Cold Start" [17233ms]
├─ "Process Initialization" [181ms] T+0.00s
│ ── ~16s idle, no spans ──
└─ "JS Bundle Execution" [1037ms] T+16.19s
Values range from ~4s to ~56s but never above 60s, matching the 1-minute cap in
AppStartMetrics. This creates wrong values forp95/avgcold start on Android.Looks like
appLaunchedInForegroundis only evaluated inonActivityCreated(), which fires when the user finally opens the app.Related fixes I checked don't cover this:
The iOS equivalent was fixed by dropping pre-warmed starts: getsentry/sentry-cocoa#1896