UserAgent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_11_4) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/52.0.2743.82 Safari/537.36

Steps to reproduce the problem:
I've noticed the following results in a large, rapid-fire amount of dataavailable events being raised:

1. Get a tabCapture stream (https://developer.chrome.com/extensions/tabCapture).
2. Use getUserMedia to get a microphone stream and add its tracks to the tabCapture video stream.
3. Record this stream via MediaRecorder using a timeslice (in my case 500ms).
4. Have a long period of inactivity on the recorded tab (either switch to another tab not being recorded or remain still - about 20s-1m has been sufficient for me).
5. Try to scroll around again and perform some activity on tab after period of inactivity.

What is the expected behavior?
Periods of idle behavior and then activity under a tabCapture stream recorded via the MediaRecorder does not result in a Chrome extension crash (my guess is does not result in a large spike in memory usage causing a Chrome extension crash).

What went wrong?
This ends up crashing our Chrome extension. I tried also recording via another mechanism (by setting up an RTC peer on our server and pulling off that stream from the backend) and there's no crash.

My intuition tells me that the MediaRecorder does not receive data from the tabCapture stream after some inactivity and then all of a sudden when there's activity again the tabCapture stream starts sending a ton of data to the MediaRecorder from the idle period, which causes it to consume a large amount of memory and then crash the extension. I tried combatting this by putting in a regular poll of calling MediaRecorder.requestData every 5 seconds, but this does not seem to solve the crash.

Did this work before? N/A 

Chrome version: 52.0.2743.82  Channel: stable
OS Version: OS X 10.11.4
Flash Version: Shockwave Flash 22.0 r0