Skip to main content

RUM Crashes

The Crashes tab groups individual crash events into deduplicated crash groups so you can prioritize by impact instead of triaging one crash at a time.

Crash list with JS Error and Native filter chips above Title, Kind, Type, Occurrences, Sessions, Users, App Version, and Last Seen columns


Crash List

Filter chips at the top let you toggle between crash Kind values; the table below lists every crash group ranked by occurrence.

Table Columns

ColumnDescription
TitleThe crash's grouping title (exception name or synthetic label)
KindHigh-level classification badge, either Native or JS Error
TypeMore specific classification, for example native_signal, mach, jvm_exception, jvm_crash, exit_self, excessive_resources, user_requested, unknown
OccurrencesTotal times this crash group has been seen
SessionsNumber of distinct sessions affected
UsersNumber of distinct users affected
App VersionVersion most recently associated with the group
Last SeenWhen the crash group was last observed

Sort by Occurrences to prioritize the crashes affecting the most sessions first.

Native is a crash the platform reported; JS Error is anything above the platform layer, which on Flutter means Dart and app-level crashes. The badge name is shared across RUM platforms - no JavaScript is involved.


Crash Details

Click a crash group to open the detail view, scoped to a specific occurrence (pick a different occurrence from the dropdown at the top).

Summary Stats

Crash title, Kind/Type badges, and Total Occurrences/Affected Users/Affected Sessions stat cards

Total Occurrences, Affected Users, and Affected Sessions for the selected crash group and time range.

Occurrence Breakdown

Occurrences by App Version, OS Version, and Device Model charts

Three charts pair a multi-line time series with a Name / Min / Mean / Max table, showing how many occurrences hit each App Version, OS Version, and Device Model:

  • Occurrences by App Version
  • Occurrences by OS Version
  • Occurrences by Device Model

Attributes

Attributes panel showing app., crash., and other grouped attributes

A scrollable panel with the full attribute set captured at crash time. A small app.* group sits above a single large crash.* block - the raw crash report, flattened - followed by the same resource attributes you see on other detail pages:

GroupExample attributes
app.*app.build, app.bundle_id, app.launch_id, app.version
crash.* (app state)crash.app_id, crash.app_version, crash.app_in_foreground, crash.app_start_time, crash.app_sessions_since_launch
crash.* (binary/build)crash.binary_cpu_type, crash.build_configuration, crash.build_type
crash.* (process/report)crash.pid, crash.parent_pid, crash.process_name, crash.report_id, crash.report_type
crash.* (signal/memory/threads)crash.signal, crash.signal_code, crash.memory_usable_bytes, crash.thread_count
crash.* (type/timing)crash.type, crash.timestamp, crash.time_zone, crash.translated
device.*device.model.name, device.locale, device.timezone, device.is_physical
host.*host.arch, host.name, host.os.name, host.processors
os.*os.name, os.version, os.build, os.type
process.*process.executable.name, process.num_threads, process.runtime.name, process.runtime.version
service.* / session.*service.name, service.version, session.id, session.start_time

Use Search attributes in the sidebar to jump straight to a specific key instead of scrolling.

Stack Trace and Translated Report

Stack Trace (Native) frames and Translated Report text block

The Stack Trace (Native) panel shows the raw symbolicated frames. Below it, Translated Report renders the full platform crash report (process info, exception type and codes, thread and register state, binary images) as plain text, with a Copy button so you can paste it straight into an issue tracker or share it with a native engineer.

Frames only resolve to readable file, line, and symbol names once you've uploaded the matching debug artifacts for that build (dSYM, Dart symbols, NDK symbols, ProGuard mappings) - see scout artifacts upload. Without an uploaded artifact, frames show raw addresses or obfuscated names instead.

Breadcrumbs table with Time, Type, and Message columns

A chronological table (Time, Type, Message) of what the user was doing right before the crash: session start, screen navigation, taps, long tasks, and UI hangs. Use it to reproduce the exact sequence that triggered the crash.

Affected Devices and Affected Users

Affected Devices and Affected Users tables

  • Affected Devices - device model, architecture, OS, app version, and occurrence count
  • Affected Users - user identifier, session count, crash count, and last hit

Use View Session Timeline to jump from a specific occurrence into the full session where it happened.


Use Cases

Triaging After a Release

  1. Filter by the new App Version
  2. Sort the Crash List by Occurrences
  3. Open the top crash group and check Occurrences by App Version to confirm it's new (or worse) in this release
  4. Use the Translated Report and Breadcrumbs to hand a native engineer a reproducible sequence

Verifying a Fix

  1. Open the crash group you fixed
  2. Filter to the version containing the fix
  3. Confirm Occurrences by App Version shows zero (or near-zero) for that version

Scoping a Device-Specific Crash

  1. Sort the Crash List by Occurrences
  2. Open a crash group and check Occurrences by Device Model - a spike concentrated on one model points at a device-specific bug (memory limits, OS fork behavior)

  • Getting Started - Interface layout and shared filters
  • Overview - Crash trends alongside sessions and ANR
  • Errors - Non-fatal errors, for issues that don't crash the app
  • Sessions - Full session timeline for a specific occurrence
  • traceX - Distributed tracing explorer
  • logX - Log explorer with trace correlation
Was this page helpful?