Last Comment Bug 815829 - Add MEMORY_VSIZE to telemetry
: Add MEMORY_VSIZE to telemetry
Product: Toolkit
Classification: Components
Component: Telemetry (show other bugs)
: unspecified
: x86_64 Linux
-- normal (vote)
: mozilla20
Assigned To: Nicholas Nethercote [:njn]
: Georg Fritzsche [:gfritzsche]
Depends on:
  Show dependency treegraph
Reported: 2012-11-27 15:06 PST by Nicholas Nethercote [:njn]
Modified: 2012-11-29 06:49 PST (History)
1 user (show)
See Also:
Crash Signature:
QA Whiteboard:
Iteration: ---
Points: ---
Has Regression Range: ---
Has STR: ---

Add MEMORY_VSIZE to telemetry. (2.05 KB, patch)
2012-11-27 23:06 PST, Nicholas Nethercote [:njn]
justin.lebar+bug: review+
Details | Diff | Splinter Review

Description User image Nicholas Nethercote [:njn] 2012-11-27 15:06:57 PST
MEMORY_VSIZE is easy to measure and could provide useful data in understanding whether Win64 builds would be helpful.
Comment 1 User image Nicholas Nethercote [:njn] 2012-11-27 23:06:35 PST
Created attachment 686002 [details] [diff] [review]
Add MEMORY_VSIZE to telemetry.

Simple patch.  I tested it via about:telemetry (which is now built-in!).
Comment 2 User image Justin Lebar (not reading bugmail) 2012-11-28 07:41:31 PST
Comment on attachment 686002 [details] [diff] [review]
Add MEMORY_VSIZE to telemetry.

Can you make the upper bound a bit bigger than 4GB?  For our 64-bit discussions I'd be curious to know how many people on 64-bit platforms have vsize > 4gb.

Also, I've concluded that it was probably a mistake to use 50 buckets in memory_resident.  (We can't change it now without messing up historical comparisons, which is why I haven't touched it.)  The telemetry front-end does a poor job of smoothing between buckets (for example when calculating averages), so the net result is that we have basically zero insight into what's happening between the buckets.  As a result, I'd be in favor of doubling or quadrupling the number of buckets here, to 100 or 200.

The downside of having so many buckets is that pings potentially get larger.  But I don't think that's a big deal.
Comment 3 User image Nicholas Nethercote [:njn] 2012-11-28 14:29:55 PST
Comment 4 User image Ed Morley [:emorley] 2012-11-29 06:49:02 PST

Note You need to log in before you can comment on or make changes to this bug.