As a Cartographers team member, I want to understand why half the sessions in Data Dog don't have a user ID recorded, so that we can make progress towards using the user ID to track new and returning users.
Notes
Details of Wes's discoveries so far is in this ticket (links to comment)
Wes has a support thread going with JT Skidmore (DD support); Wes & Daniel are meeting JT on Nov 5.
Possible tasks:
Determine whether DD is configured in Staging to run a 100% sample rate. (Prod has had sampling issues after the Identity team reduced sampling to 20% and it bled over to health apps.)
Vets API: Verify whether authenticated users can have no accountUuid (or API will return nil even if they do), and if so, how often could that occur
When logged in to a staging session, look at RUM cookies to see what data is stored there. There is a bug where DDG stores certain configuration in a cookie (e.g. sampleRate), so one SPAs DDG config can contaminate other SPAs behavior. It is believed that the bug shouldn’t affect user id due to our datadogRum.setUser( call, but we should investigate.
Acceptance criteria
UserID is recorded for 90%+ of sessions, if possible
Description
User story
As a Cartographers team member, I want to understand why half the sessions in Data Dog don't have a user ID recorded, so that we can make progress towards using the user ID to track new and returning users.
Notes
Possible tasks:
sampleRate
), so one SPAs DDG config can contaminate other SPAs behavior. It is believed that the bug shouldn’t affect user id due to ourdatadogRum.setUser(
call, but we should investigate.Acceptance criteria