1. Overview
Crash Analysis catches and analyzes crashes that occur in the main and child processes of an application during runtime. It processes crash logs to provide developers with error trends, error distribution, an error list, and error details for a selected time frame. This information helps you quickly analyze the root causes of application crashes.
2. Terms
Error count: The number of times a specific type of error occurs within a time range. This is equivalent to the number of logs reported by the software development kit (SDK).
Error rate: Error Rate = Error Count / Number of Application Starts
Affected users: The number of unique devices that experienced a specific type of error within a time range. This value is calculated by counting the unique devices from all relevant error logs.
Affected user percentage: Affected User Percentage = Number of Affected Users / Number of Active Users
3. Feature introduction
3.1 Error trends
This section displays trend charts for the error count, error rate, number of affected users, and affected user percentage based on the selected filters and time range. You can compare data across different application versions.
Exception type
On Android, crashes are categorized as Java crashes and Native crashes.
On iOS, the system catches crashes that occur in the Objective-C or Swift layer.
Time range
You can query data from the last 90 days.
Dimension filters
You can filter by dimensions such as application version, device model, operating system, carrier, channel, and region.
Advanced features
New exceptions: These are exceptions that are assigned an error ID for the first time within the selected time range and filter dimensions. The error ID is the unique identifier for an error summary.
Block system-level errors: If an error stack contains no application-specific code, it is classified as a system-level error. You can block system-level errors to focus on analyzing application-specific errors.
View only HarmonyOS devices: Select this option to view only errors or exceptions that occurred on HarmonyOS devices. Note that system-level errors on HarmonyOS are not yet supported.
3.2 Error distribution
This section displays distribution charts based on the selected time range and filters. The data is categorized by eight dimensions: application version, device model, operating system, carrier, channel, region, crash page, and runtime duration.
Top 5: A ranking of the top five items by error count.
Details: The breakdown of error counts and user counts for each property value within a dimension. You can filter by specific properties in the table header.
3.3 Error list
This section lists errors based on the selected time range and filters. The list is generated by aggregating all crash logs by crash summary.
Error ranking: By default, errors are sorted in descending order by crash count to help you quickly identify the most frequent and critical errors. You can also sort by the number of affected users to find the errors with the widest impact.
Error ID: The unique identifier for an error summary.
Version range: The range of application versions in which the crash occurred, based on the current filters.
Resolution status: The status of the error. Supported statuses include Unresolved, In Progress, Ignored, and Resolved.
Alert settings: You can set up alerts for individual errors. Alerts added here can be modified and configured in the Monitoring and Alerting section.
Tags: You can add tags to error summaries.
Export: You can export the currently displayed error list. You can export up to 1,000 entries at a time.
3.4 Error details
This section aggregates error logs with the same error summary and displays detailed information. This helps you quickly locate and resolve issues.