Telemetry probes yielding relevant data

Jonathan Protzenko jonathan.protzenko at gmail.com
Wed Oct 5 08:32:27 UTC 2011


Hi folks,

Just a followup on the Telemetry work I did this summer. The telemetry 
opt-in patch landed right before the merge, which means there now is an 
opt-in bar on Thunderbird 10 and Thunderbird 9 (daily / c-c + earlybird 
/ c-a). We now have ~80 people sending Telemetry data which you can read 
at 
<https://metrics.mozilla.com/pentaho/content/pentaho-cdf-dd/Render?solution=metrics&path=telemetry/&file=TelemetryHistogram.wcdf> 
assuming you have the right credentials.

The three probes we have are:
- THUNDERBIRD_GLODA_SIZE_MB : many people have a size of 1MB or less, 
which tends to make me think these people have Gloda disabled. I wonder 
how much of a correlation there is between being a tester and disabling 
Gloda; there's only one pathological user with a size that's greated 
than 1GB.
- THUNDERBIRD_CONVERSATIONS_TIME_TO_2ND_GLODA_QUERY : very few people 
reporting on that one, I think the Telemetry reporting code is not in 
the currently public version of Conversations, so I guess I need to make 
a new release
- THUNDERBIRD_INDEXING_RATE_MSG_PER_S : the distribution is pretty 
smooth, and the mean seems to be around 30 messages per sec, although 
there's an alarmingly high number of people indexing 100 messages per 
sec; maybe we need to make the scale bigger (do they have SSDs?).

I've asked the telemetry people if there's a way we can display the mean 
value, which would be especially important for the gloda database size 
probe, for instance.

jonathan



More information about the tb-planning mailing list