telemetry data on sync success rate
rfkelly at mozilla.com
Fri Sep 26 10:11:59 PDT 2014
On 27/09/2014 2:53 AM, James Bonacci wrote:
> Well, this sure looks good.
Heh, it's certainly re-assuring as a sanity-check. But I did very
deliberately put in this disclaimer:
> I'm not entirely sure how to interpret the measures here, and we
> probably should be measuring something like "user saw an error bar" more
For example, I don't know whether an early FxA failure would error out
before triggering WEAVE_START_COUNT.
I'll poke at the code a bit and file some bugs for more deliberate
measurement of error conditions here.
> But it's interesting to compare:
> For "release32" I see:
> 36.05M started, 35.62M completed successfully (98.8%)
> For "nightly34" I see:
> 5.26M started, 5.18M completed successfully (98.4%)
> Sync-dev mailing list
> Sync-dev at mozilla.org
More information about the Sync-dev