Add build and launch telemetry #594
Conversation
|
The latest updates on your projects. Learn more about Vercel for Git ↗︎
|
| @@ -100,6 +101,11 @@ export class DeviceSession implements Disposable { | |||
| } | |||
|
|
|||
| private async launchApp() { | |||
| const launchReequestTime = Date.now(); | |||
There was a problem hiding this comment.
typo:
launchReequestTime -> launchRequestTime
| waitForAppReady.then(() => clearTimeout(reportWaitingStuck)); | ||
| } | ||
|
|
||
| const [previewUrl] = await Promise.all([this.device.startPreview(), waitForAppReady]); |
There was a problem hiding this comment.
Shouldn't we wait for waitForAppReady once?
| waitForAppReady.then(() => clearTimeout(reportWaitingStuck)); | |
| } | |
| const [previewUrl] = await Promise.all([this.device.startPreview(), waitForAppReady]); | |
| } | |
| const [previewUrl] = await Promise.all([this.device.startPreview(), waitForAppReady]); | |
| if (shouldWaitForAppLaunch) { | |
| clearTimeout(reportWaitingStuck); | |
| } |
There was a problem hiding this comment.
I like my approach moare as it doesn't mix the reporting code with the existing logic. Whatever is placed inside that if block I added, is just for reporting and can be deleted. Otherwise I'd have to keep the reportWaitingStuck in main scope, while we'd need to have shouldWaitForAppLaunch checks both before and after the await.
On top of that this isn't really an equivalent, as the await is for both startPreview and app ready event.
| const [previewUrl] = await Promise.all([this.device.startPreview(), waitForAppReady]); | ||
| Logger.debug("App and preview ready, moving on..."); | ||
| this.eventDelegate.onStateChange(StartupMessage.AttachingDebugger); | ||
| await this.startDebugger(); | ||
|
|
||
| const launchDurationSec = (Date.now() - launchReequestTime) / 1000; |
There was a problem hiding this comment.
typo:
launchReequestTime -> launchRequestTime
This PR adds more loggind and telemetry around build and launch process.
Building and launching the app is one of the most fragile element of the IDE pipeline and we get numerous reports around that. While there are a few known issues around it that we are fixing and investigating. I figured it'd be good to also add telemetry around that such that we can track how it works for a larger group of people and also be able to detect regressions.
Test plan
Just run the example app with clean build and make sure nothing breaks.