diff options
author | jar@chromium.org <jar@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98> | 2009-02-09 05:49:22 +0000 |
---|---|---|
committer | jar@chromium.org <jar@chromium.org@0039d316-1c4b-4281-b951-d872f2087c98> | 2009-02-09 05:49:22 +0000 |
commit | 54131d25aa7e887f89ba25851b983150cd1384a8 (patch) | |
tree | 98772c3da4a2d43dd69f1c125a855743289bfe33 /base | |
parent | 005403a5916926d45fa20a73e13222eb2be56e05 (diff) | |
download | chromium_src-54131d25aa7e887f89ba25851b983150cd1384a8.zip chromium_src-54131d25aa7e887f89ba25851b983150cd1384a8.tar.gz chromium_src-54131d25aa7e887f89ba25851b983150cd1384a8.tar.bz2 |
Force all stability stats into "initial log" (don't use ongoing log)
The goal of this checkin is to effectively roll-back our stats gathering
methodology for stability stats to ONLY use the initial log. This has
some negative consequences (time lag, and bias against data from happy
ongoing users). We're doing this because there appears to be a BIG
survivability differnce between initial and ongoing logs. Perhaps we
are even getting multiple recordings of data from initial logs.
Until this is corrected, it is not safe to send part of the stability
info (crashes) only in the initial logs, while we send pageloads in
part via ongoing logs.
To absolutely minimize risk, I decided to leave the inter-log delay
at 20 minutes, so that there is no chance of changing the log saving
resource utilization (i.e., going over the limits). It was at 5
when we started the process, but since it has no impact on the initial
log, it won't harm the stability stats, and it may reduce load on
the front end log saving servers.
bug=1564631
r=huanr,evanm
Review URL: http://codereview.chromium.org/21160
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@9385 0039d316-1c4b-4281-b951-d872f2087c98
Diffstat (limited to 'base')
0 files changed, 0 insertions, 0 deletions