[Due for payment 2026-06-01] [$250] Web app crashes intermittently with no console errors

[MelvinBot]

Description

The web app is crashing intermittently with no errors visible in the JS console after the crash occurs. This makes it difficult to debug since the console state is lost on crash.

Most likely this crash https://expensify.sentry.io/issues/APP-2BN

Steps to reproduce

Unknown — needs reproduction. The reporter experienced multiple crashes on the web app but was not able to capture console output at the time of the crash.

Expected behavior

The web app should not crash unexpectedly.

Actual behavior

The web app crashes intermittently. No JS console errors are visible after the crash (likely because the page reloads and console is cleared).

Additional context

• Consider leveraging Sentry for better crash reporting so we don't rely on users having the console open
• Slack thread: https://expensify.slack.com/archives/C05LX9D6E07/p1778782686789009

Upwork Automation - Do Not Edit

• Upwork Job URL: https://www.upwork.com/jobs/~022055203627066608029
• Upwork Job ID: 2055203627066608029
• Last Price Increase: 2026-05-15

Issue Owner

Current Issue Owner: @eVoloshchak

[exfy-zapier]

This has been labelled "Needs Reproduction". Follow the steps here: https://stackoverflowteams.com/c/expensify/questions/16989

[melvin-bot]

📣 @exfy-zapier! 📣
Hey, it seems we don’t have your contributor details yet! You'll only have to do this once, and this is how we'll hire you on Upwork.
Please follow these steps:

1. Make sure you've read and understood the contributing guidelines.
2. Get the email address used to login to your Expensify account. If you don't already have an Expensify account, create one here. If you have multiple accounts (e.g. one for testing), please use your main account email.
3. Get the link to your Upwork profile. It's necessary because we only pay via Upwork. You can access it by logging in, and then clicking on your name. It'll look like this. If you don't already have an account, sign up for one here.
4. Copy the format below and paste it in a comment on this issue. Replace the placeholder text with your actual details.
   
   Format:

Contributor details
Your Expensify account email: <REPLACE EMAIL HERE>
Upwork Profile Link: <REPLACE LINK HERE>

[MelvinBot]

Issue Analysis

⚠️ Medium confidence — The "Aw, Snap!" crash is a Chrome renderer/tab crash (not a JS exception), most likely caused by out-of-memory (OOM) conditions. Without reproducible steps, the exact trigger is uncertain, but the architectural patterns below are the most probable contributors.

Root Cause: Chrome's "Aw, Snap!" page indicates the renderer process crashed — this is almost always an OOM condition on web. The App loads entire Onyx collections (all reports, transactions, report actions) into memory at startup and keeps them resident. Users with large accounts can exhaust Chrome's JS heap limit (~2GB desktop, ~500MB mobile), causing a full tab crash. Because this is a process-level crash (not a JS error), no console output or error boundary can capture it.

Why existing crash reporting doesn't help:

1. The ErrorBoundary catches React errors and unhandled promise rejections, but OOM tab crashes kill the process before any JS can run
2. Memory telemetry sends breadcrumbs to Sentry every 2 minutes, but is passive — it observes memory usage but takes no preventive action when thresholds are exceeded
3. No window.onerror handler exists for synchronous global errors

Key contributing patterns:

• Multiple locations use waitForCollectionCallback: true to load ALL reports into memory simultaneously (tracked as a known issue in Remove Onyx.connect() for the key: ONYXKEYS.COLLECTION.REPORT in src/libs/OptionsListUtils.ts #66378)
• No web-specific memory management or data eviction strategy
• Image/attachment data (including base64-encoded images) stored in Onyx without size limits

Investigation details

Memory architecture gaps

The app has memory monitoring via sendMemoryContext with thresholds at 70% (warning) and 85% (critical) of jsHeapSizeLimit, but this only logs to Sentry — there is no auto-reload, data eviction, or garbage collection trigger when memory is critical.

Why console is empty after crash

Chrome's "Aw, Snap!" is a renderer process crash. The DevTools console runs in the same renderer process, so it is destroyed along with all JS state. This is fundamentally different from a JS exception — no error handler can catch it because the process is already dead.

Alternative explanations considered

• WebGL/Canvas crash: The app uses canvas for text measurement and avatar processing, but these are lightweight operations unlikely to crash the renderer
• Service Worker issue: SW registration has error handling but no retry logic; however, SW crashes would not cause "Aw, Snap!"
• Browser extension conflict: Possible but unlikely to be the primary cause given the memory architecture concerns

Possibly related issues

• Remove Onyx.connect() for the key: ONYXKEYS.COLLECTION.REPORT in src/libs/OptionsListUtils.ts #66378 — Onyx.connect() refactoring to avoid loading all reports into memory (directly related architectural debt)

Recommended next steps:

1. Check Sentry for memory breadcrumbs from the reporter's session — the 2-minute memory tracking may show a ramp-up before the crash
2. Reproduce by loading a large account (100+ reports) on web and monitoring performance.memory.usedJSHeapSize in DevTools
3. Short-term mitigation: Add an active memory pressure response (e.g., tab reload with state recovery when memory exceeds 85%)
4. Long-term fix: Complete the Onyx.connect() refactoring (Remove Onyx.connect() for the key: ONYXKEYS.COLLECTION.REPORT in src/libs/OptionsListUtils.ts #66378) to avoid keeping all collections in memory

---

Next Steps: Reply with @MelvinBot implement this to create a draft PR, or @MelvinBot <your feedback> to refine this analysis.

[melvin-bot]

Triggered auto assignment to Contributor-plus team member for initial proposal review - @eVoloshchak (External)

[melvin-bot]

Job added to Upwork: https://www.upwork.com/jobs/~022055203627066608029

[mountiny]

Looks like its this Sentry crash https://expensify.sentry.io/issues/APP-2BN