Telemetry/2013-02-07
From MozillaWiki
Metrics/Telemetry 2013-02-07
Attendees: Daniel Einspanjer, Annie Elliott, Taras Glek, Lawrence Mandel
Status
BIV-70 (bug 776714) and BIV-69 (bug 776716) Adding descriptions to Telemetry dashboard
- done
BIV-224 (bug 810034) Telemetry Dashboard session timeout is too quick
- Still testing
METRICS-1356 (bug 778809) Simple measures
- Taras missed sending e-mail with target measures, measures now added in Jira and bug
- This will be implemented in ETL
BIV-114 (bug 707280) Need simple queries
- Harsha looking at initial work to prototype it out, sidetracked with spike in telemetry submissions METRICS-1366
METRICS-1366 Telemetry submission spike has been taking a lot of time
- 5 mil to 21 mil submissions in a single day is bad for a number of reasons
- not a steady trend of submissions wondering how valid is data coming in, no evidence of this type of fluctuation within a weekday
- all channels went up but release channel went up 12 million as it's a bigger channel
- not equipped to handle this amount of traffic, strain put on resources (elastic search)
- every few days see a spike of submissions
- haven't found a pattern
- Harsha working full time to attempt to fingerprint the documents looking for indication of source
- comparison of idle daily vs saved sessions and trend is holding stable
- can deal with capacity side but how can we trust this data if it's behaving in a way that we don't understand?
- would like to get someone on dev side to look at submissions from client side to try and find more useful information
- Taras: there have been no changes in date range that Daniel provided, nothing else to act on without more guidance from metrics
- Taras requests that Metrics look at the IPs from the submissions
METRICS-1363 (bug 821403) Make raw shutdown times available
- No progress
Process
Lawrence: Engineering didn't know that all work was stopped due to the Telemetry submission investigation. Annie: We should have communicated the work stoppage. Telemetry has 1 FTE allocated, which is why all work stopped while Harsha was 100% on the submission spike investigation.
Iteration 2 planning
- All work is stopped until the submission spike has been resolved in METRICS-1366
- Taras requests daily updates on this issue until it is resolved, please cc Taras Glek, Vladan Djeric, Nathan Froyd
- BIV-110 not ready to start
- BIV-85 and BIV-224 should be done in the next two weeks