Skip to main content
After errors are reported, you can view Issues in the Error Tracking module. In RUM, an Issue is a group of similar errors, typically related to the same bug.
Issue List
For detailed grouping rules, please refer to Error Grouping.

Issue information overview

Issue Information Overview
Each entry listed in the Issue browser contains the following information:
The Issue list can be sorted by priority, and the Priority filter is included in the default filters. The accent bar to the left of each Issue’s error message reflects its priority (Critical red, Warning orange, Info light orange; gray when no priority data exists). Priority is determined by alert rules or the scoring system and cannot be edited manually; it is also shown at the top of the Issue detail page.

Issue status

An Issue is For Review, Reviewed, Resolved, or Ignored. Resolved has two modes, “Reopen on recurrence” and “Reopen in a new version”, which are submenu items under Resolved in the status dropdown. For the transition rules, see Issue status.

Filtering and sorting

The timeline in the upper right corner of the browser allows you to display Issues that had errors during the selected time period. You can:
  • Select a preset range from the dropdown menu
  • Directly modify the time
  • Enter natural language for filtering

Export issues

The CSV export button at the top right of the Issue list exports data that matches the list’s current filters and sort order. After clicking the CSV button, you can choose what to include in the export popover:
Error samples only exist within the error data retention window; older Issues export with empty sample columns.
The export always includes 18 Issue fields: Issue ID, Issue URL, Application, Service, Error type, Error message, Status, Severity, Is crash, Error count, Affected sessions, First seen, First seen version, Last seen, Last seen version, Versions, Suspected cause, and Resolved at. If you turn on “Include error samples”, the sample columns are appended.
  • A single export is capped at 100 rows. If more match, narrow the time range or add filters.
  • Time columns (First seen / Last seen / Resolved at / sample time) show the timezone in the header, for example First seen (Asia/Shanghai). Cells hold YYYY-MM-DD HH:mm:ss text in that timezone, so spreadsheets sort them as dates.
  • To prevent formula injection, cells starting with =, +, -, or @ are prefixed with a quote.
  • Sample stacks are truncated to the first 20 lines. Full stacks are available in the Issue detail (where they are symbolicated per platform); the export contains the raw frames as reported, without symbolication.

Error cause classification

Flashduty labels each Issue with a likely error cause when it is created, to narrow down where to look.
Hover over an error cause to see the root cause and fix suggestions inferred by AI.

How classification works

The system first classifies by error type and message. For example, a request with an HTTP 4xx or 5xx status is API Request Failed, and a TypeError on a null value is Illegal Object Access. When it can’t tell, AI uses the error message, stack trace, and platform to give a classification and a cause explanation of up to 100 words, shown as a hover tooltip on the Issue card.

Regression

A resolved Issue that occurs again reopens and is marked as a regression. With “Reopen on recurrence”, any recurrence reopens it, including in the same version; with “Reopen in a new version”, only a newer version does. See Issue status.

Issue details

The Issue list supports two viewing modes: sidebar mode and fullscreen mode. By default, the details panel opens as a sidebar. You can click the expand button to switch to fullscreen mode for a more spacious view and more complete data display. Click any Issue to open the details panel and view more information.
Issue Details
The top of the panel displays basic Issue information, such as status and error cause. You can also understand the Issue lifecycle: first and last occurrence dates, duration, and error occurrence count over time (aggregated by time granularity).
In the label distribution section, you can view the proportion of different labels under this Issue across various dimensions, quickly determining the impact scope and helping locate root causes.Currently supports labels like view_name, browser_name, version, env, and so on.
By default, displays the most recent error information reported during the Issue’s occurrence as an error sample. You can also switch through the navigation bar.Each native crash sample shows whether it is symbolicated. If the latest sample can’t be symbolicated, the detail switches to an earlier sample that can. You can switch back to the latest one.When a crash is known so far only from the beacon the app sent as it died, and its full report has not arrived, the sample area shows Crash details pending with a one-line explanation instead of a stack. The stack and the rest of the details upload when that user opens the app again, and replace this sample.
View error context and stack information. If the corresponding SourceMap, Android mapping file, iOS dSYM file, or Flutter symbol file has been uploaded, you can see mapped original source code locations and code snippets. For Web, WeChat Mini Program and HarmonyOS errors, code snippet lines longer than 480 characters are clipped, and each cut end is marked with ….
Error Stack
You can view uploaded source code information in “Application Management” - “Source Code Management”. See Source Mapping for details.
The error stack display for native platforms (Android/iOS) is specifically designed for mobile characteristics, providing the following capabilities:Pretty / Raw Mode Toggle
  • Pretty Mode: Displays structured stack information, automatically distinguishing between app frames and third-party frames. Third-party frames are collapsed by default, highlighting your own code
  • Raw Mode: Displays raw stack text for easy copying and analysis in external tools
Symbolication StatusIf you uploaded symbol files (Android mapping or iOS dSYM) for that version, obfuscated addresses in the stack are restored to function names, file names, and line numbers. When not symbolicated, the system prompts you to upload symbol files with direct links to “Source Code Management” and the upload tool.Symbolication results are cached. After you upload symbols for a crash you’ve already seen, click Re-symbolicate in the notice bar or in the “Symbols required by this crash” section. This bypasses the cache, reruns symbolication (including re-fetching dSYMs), and refreshes the stack. You don’t need to wait for a new occurrence.Thread Stacks (Threads)Native crashes typically involve multiple threads. The thread panel displays stack information for all threads at the time of the crash, supporting:
  • Viewing total thread count and currently displayed threads
  • Expanding/collapsing all threads
  • Independent separation of app frames and third-party frames within each thread, with third-party frames expandable on demand
  • The crashing thread’s stack is displayed first
Binary Images (iOS)For iOS crashes, you can also view the list of Binary Images loaded at the time of the crash, including image names, address ranges, and UUIDs, for offline symbolication analysis.Symbols Required by This Crash (Non-System)For native crashes such as iOS and Electron, including Flutter’s iOS native crashes, the non-system binary images are listed above the stack. Each shows its name, UUID, and symbolication status. The list stays expanded while anything is missing. You can:
  • copy the UUID, or click Look up in symbol list to open Source Mapping filtered by that UUID;
  • click Upload on an unparsed image to upload its symbols;
  • click Re-symbolicate after uploading to refresh the stack.
Download crash logThe stack area can export this crash as a plain-text log file (Download crash log) containing the summary, the crashing stack, every thread’s stack and the binary images. Every stack that was symbolicated is repeated in its original text in the Original Stacks section at the end of the file.Flutter SupportFlutter native crashes (with source_type of ndk, android, or ios) carry thread stacks and Binary Images, and are rendered with the same native view described above as Android/iOS native crashes. Dart exceptions are symbolicated by matching the build_id in the stack against uploaded Flutter symbol files.React Native SupportReact Native works like Flutter. JS errors render as Web stacks and are symbolicated with uploaded JS sourcemaps. Native crashes (source_type of ios, android, or ndk) carry thread stacks and Binary Images. They render in the native stack view above and are symbolicated with dSYM / NDK symbol files, like standalone Android/iOS crashes. Sourcemap lookup for JS errors is scoped to the react-native type and to the platform from the error’s OS (os_name). iOS and Android bundles can share a file name, and platform scoping keeps them from matching each other’s file.Electron SupportElectron errors are routed by type. Main-process and renderer JavaScript errors are V8 stacks and render like browser errors in the Web stack view. Uploading a source map restores the original source. Minidump native crashes are address-style stacks with threads and binary_images, shown in the native stack view above. Uploading matching Breakpad symbols restores function names, file names, and line numbers. See Electron error symbolication.For detailed symbol file upload procedures, see Source Mapping.
View the total number of Session events for the current error sample, as well as user resource access and operations before and after the error occurred.
Currently displays up to 20 context items including the current Error event. You can view more log information in the Session Viewer module later.
If the session has replay data, click View Replay to open the session replay. It shows what the user did when the error occurred.
Error events carry a series of attributes when reported. You can view current Session, view, user, and other information in the attributes section to help troubleshoot issues.For Electron errors, the attributes panel also shows a virtual process attribute (main process / renderer process) in the Other attribute group. It tells you which runtime the error came from. Other errors don’t show it.

Header actions

The top right of the Issue details panel has two AI SRE buttons. They show only when AI SRE is enabled for the account: Auto-analyze is triggered by incidents, so the application must be able to raise them first. When you click the button, RUM reads the application’s alert settings. Only if alerting is on, delivery is Flashduty channel, and at least one channel is selected does it go straight to the create form, with channels already picked from the application’s alert settings; confirm and click Create. Otherwise it shows “This application has no On-call alerting yet. Issues must raise incidents to trigger auto-analysis, so set up alerting first.” and jumps to the application’s Alert settings tab. For the template and form fields, see Automations.

Issue alerts

In Application management, edit the application, turn on Alerts, and select a channel or Webhook. For setup, see Issue alerts.

Best practices

  • Configure Source Mapping: Facilitates problem localization in production environments
  • Add User Information: Configure user-related information to provide better error context
  • Set Alert Policies: Configure appropriate channels and escalation rules for errors
  • Regular Error Analysis: Regularly check error reports to discover potential issues
  • Team Collaboration: Use team ownership features to ensure issues are quickly assigned to relevant teams
  • Monitor Regressions: Closely monitor potential regressions of resolved issues

Next steps

Source Mapping

Configure source mapping

Error Grouping

Learn about grouping mechanisms

Issue Status

Manage Issue status