Trace Network Quality Estimator |
||||
Issue descriptionWow, NetworkQualityEstimator looks useful! I wonder if it might be useful for tracing? If a netlog event looks long, it might help speed up debugging to know that RTT was high or throughput was low. Eventually, these estimations might even be useful for metrics. Trace viewer would probably want to visualize this information as sparklines, not object snapshots. I'd need to dig to see how sparklines like memory usage are traced. Might be related to tracing UMA histograms.
,
Apr 20 2016
The RTT and throughput estimates might be useful to have in the trace log, in addition to trace events covering NQE's methods, and in addition to having the estimates in UMA. No particular use cases need this yet. This came up in a discussion about parameterizing normalization functions for TBMv2 loading metrics as a nice-to-have-eventually. ChartTrack in catapult looks like the appropriate UI base class.
,
Feb 25 2017
,
Mar 14 2017
,
Nov 8 2017
Refreshed during triage.
,
Dec 27 2017
Refreshed during triage.
,
Feb 13 2018
Refreshed during triage.
,
Apr 11 2018
Is there any use cases that need this yet? Otherwise, it might just make sense to close this bug until any use cases that need it come up.
,
Apr 18 2018
Lets reopen this if it's needed later. For now, it does not seem like NQE is consuming too many CPU cycles. |
||||
►
Sign in to add a comment |
||||
Comment 1 by tbansal@chromium.org
, Apr 20 2016