Application Not Responding (ANR)
Application Not Responding (ANR) errors, or Event Loop Stall errors, are triggered when the Electron main or renderer processes event loop is blocked for more than five seconds. The Electron SDK reports ANR errors as Sentry events and can optionally attach a stack trace of the blocking code to the ANR event.
This feature is currently in Beta. Beta features are still in progress and may have bugs. We recognize the irony.
(Available in version 4.17.0 and above)
ANR detection requires Electron v22 or higher.
ANR tracking can be individually enabled for the main and renderer processes.
For the main process:
import * as Sentry from "@sentry/electron/main";
Sentry.init({
dsn: "https://examplePublicKey@o0.ingest.sentry.io/0",
integrations: [Sentry.anrIntegration({ captureStackTrace: true })],
});
For renderer processes:
import * as Sentry from "@sentry/electron/renderer";
Sentry.init({
dsn: "https://examplePublicKey@o0.ingest.sentry.io/0",
anrDetection: { captureStackTrace: true },
});
You can pass a configuration object to both the Anr
integration and the anrDetection
renderer option to customize the ANR detection behavior.
interface Options {
/**
* Interval to send heartbeat messages to the child process.
*
* Main process: Defaults to 50ms.
* Renderer process: Defaults to 1000ms.
*/
pollInterval: number;
/**
* Threshold in milliseconds to trigger an ANR event.
*
* Defaults to 5000ms.
*/
anrThreshold: number;
/**
* Whether to capture a stack trace when the ANR event is triggered.
*
* Defaults to `false`.
*/
captureStackTrace: boolean;
}
ANR detection in the Electron main process uses a worker thread to monitor the event loop in the main app thread. The main app thread sends a heartbeat message to the ANR worker thread every 50ms. If the ANR worker does not receive a heartbeat message for 5 seconds, it triggers an ANR event. If the captureStackTrace
option is enabled, the ANR worker uses the inspector
module to capture stack traces via the v8 inspector API.
Once an ANR event is reported, the ANR worker thread exits to prevent further duplicate events and the main app thread will continue to run as usual.
Overhead from Node.js ANR tracking should be minimal. With no ANR detected, the only overhead in the main app thread is polling the ANR worker over IPC every 50ms by default. The ANR worker thread consumes around 10-20 MB of RAM to keep track of the polling. Once an ANR has been detected, the main thread is paused briefly in the debugger to capture the stack trace frames. At this point, the event loop has been blocked for seconds so the debugging overhead is negligible.
Our documentation is open source and available on GitHub. Your contributions are welcome, whether fixing a typo (drat!) or suggesting an update ("yeah, this would be better").