Re: [webrtc-stats] Is keeping stats around a memory problem?

> I do not see how I would make an application do the above for us. Wouldn't the event be sufficient for anyone to poll right away and snapshot the data?

Sorry, by "implementation" I meant the webrtc-stats implementation, meaning the browser. I'm saying "if storing historical stats takes too much memory, the browser could mitigate the problem by using hard drive space". That doesn't solve the problem though, just delays it.

I guess I'm just hoping we can avoid making major changes to the stats API to solve this corner case. Having an event for deleted objects would work, but it would be a bit of a burden both for browsers and stats clients, and as far as I know this problem is only theoretical so far.

-- 
GitHub Notification of comment by taylor-b
Please view or discuss this issue at https://github.com/w3c/webrtc-stats/issues/235#issuecomment-327940327 using your GitHub account

Received on Thursday, 7 September 2017 22:37:48 UTC