[Reporting] Investigate cause of screenshot capture timeouts on Windows

[tsullivan]

Kibana version: 6.8, 7.1

Elasticsearch version: 6.8, 7.1

Server OS version: Windows Server

Browser version:

Browser OS version:

Original install method (e.g. download page, yum, from source, etc.): download page

Describe the bug:

When a user tries to generate a report, the Reporting plugin succeeds in opening the page and scanning the DOM for visualization elements. When all visualizations have signaled rendering is done, Reporting attempts to capture a screenshot, and it times out.

Steps to reproduce:

1. So far, we have been unable to reproduce the issue in controlled environment
2. Having Cisco AMP for Endpoints and Tripwire Enterprise Agent installed might be necessary to reproduce the bug.

Expected behavior:

Screenshots (if relevant):

Errors in browser console (if relevant):
No errors logged from Puppeteer's screenshot function. The reporting job fails due to time out.

Provide logs and/or server output (if relevant):

Any additional context:

[elasticmachine]

Pinging @elastic/kibana-reporting-services (Team:Reporting Services)

[tsullivan]

For reference, this is the method that appears to be blocked: https://github.com/puppeteer/puppeteer/blob/21c2d31/docs/api.md#pagescreenshotoptions

[elasticmachine]

Pinging @elastic/kibana-app-services (Team:AppServices)

[tsullivan]

I'm still able to reproduce this on Windows 10 using a build candidate of the latest master branch.

System stats:

• 64G ram
• 12 core
• 4GB GPU

I don't understand why a decently provisioned system like this is unable to render a dashboard which Linux can, when booted on the same hardware. But I also experience general slowness with the Windows system overall, and I don't know why.

[tsullivan]

This is possibly fixed in the latest version of Kibana that uses a newer version of Chromium.exe

[tsullivan]

This seems to be no longer an issue as of #98688