Glean: Validate WorkManager implementation
Categories
(Toolkit :: Telemetry, enhancement, P1)
Tracking
()
People
(Reporter: travis_, Assigned: Dexter)
References
Details
(Whiteboard: [telemetry:mobilesdk:m8])
Make sure the WorkManager implementation fixes the sequence number holes issue we were seeing.
Assignee | ||
Comment 1•6 years ago
|
||
I would also validate:
- that the number of pings we receive from versions of glean using the workmanager is comparable with the number of pings we received before;
- that repeating Chris' analyses on the sequence numbers yields better (less to no holes) results;
- that the number of duplicate pings is comparable to the ones we were seeing before.
Reporter | ||
Comment 2•6 years ago
|
||
Just took another look at this today: query
Looks like number of pings is comparable accounting for the timespan, it may have dipped a bit due to lower MAU/DAU of the reference browser.
Repeating Chris's analysis yielded a single hole since the WorkManager changes were introduced into the reference browser, so that seems to have fixed the hole issue.
Here's the bad news, duplicates are up a lot but the duplicates also have a duplicate document_id in most cases. See here
Reporter | ||
Updated•6 years ago
|
Reporter | ||
Comment 3•6 years ago
|
||
This is going to have to wait a bit since we are about to kill all of our current data. I'd like to hang onto this but I'm going to P3 it until we have more data with the new schema.
Reporter | ||
Updated•6 years ago
|
Assignee | ||
Updated•6 years ago
|
Assignee | ||
Updated•6 years ago
|
Assignee | ||
Updated•6 years ago
|
Assignee | ||
Comment 4•5 years ago
|
||
The analysis in bug 1525603 confirmed that the "holes" problem was fixed. And also confirmed we do have a "duplicates" problem now. Closing this as FIXED, I'm filing a follow-up for the duplicates.
Description
•